IAI Accepts, published online ahead of print on 2 September 2014 Infect. Immun. doi:10.1128/IAI.02131-14 Copyright © 2014, American Society for Microbiology. All Rights Reserved.
1
The type III secretion effector NleF of enteropathogenic Escherichia coli activates NF-
2
κB early during infection
3 4 5 6
Running title: The EPEC effector NleF activates NF-ĸB
7
Mitchell A. Pallett,a Cedric Berger,a Jaclyn S. Pearson,b Elizabeth L. Hartlandb and Gad
8
Frankela#
9 10
MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences,
11
Imperial College London, UKa, Department of Microbiology and Immunology, University of
12
Melbourne at the Peter Doherty Institute for Infection and Immunity, Victoria 3010,
13
Australiab
14 15 16 17
#Corresponding author
18
Gad Frankel
19
CMBI, Flowers Building
20
Imperial College
21
London, SW7 2AZ
22
[email protected] 23
Tel: +44 20 7594 5253
1
24
Abstract
25
The enteric pathogens enteropathogenic Escherichia coli and enterohemorrhagic E. coli
26
employ a T3SS to manipulate the host inflammatory response during infection. Previously, it
27
has been reported that EPEC, in a T3SS-dependent manner, induces an early pro-
28
inflammatory response through activation of NF-κB via ERK1/2 and PKCζ. However, no
29
effector has yet been attributed to the activation of NF-κB during infection. At later time
30
points post infection NF-κB signalling is inhibited through the translocation of multiple
31
effectors including NleE and NleC. Here we report that the highly conserved non-LEE
32
encoded effector F (NleF) shows both diffuse and mitochondrial localization during ectopic
33
expression. Moreover, NleF induces nuclear translocation of NF-κB p65 and the expression
34
of IL-8 following ectopic expression and during EPEC infection. Furthermore, the pro-
35
inflammatory activity and localization of NleF was dependent on the C-terminal amino acids
36
LQCG. Whilst the C-terminal domain of NleF was previously shown to be essential for
37
interaction with caspase-4, -8 and -9; the pro-inflammatory activity of NleF was independent
38
of caspase-4, -8 or -9 interaction. In conclusion EPEC, through the T3SS-dependent
39
translocation of NleF, induces a pro-inflammatory response in an NF-κB dependent manner
40
in the early stages of infection.
2
41
Introduction
42
Enteropathogenic Escherichia coli (EPEC) and enterohaemorrhagic E. coli (EHEC) are
43
important enteric human pathogens. EPEC is a leading aetiological agent of paediatric
44
diarrhoeal disease (1); EHEC infection can lead to haemorrhagic colitis and in more severe
45
cases to kidney damage and the progression to haemolytic-uremic syndrome (HUS) (2).
46
Although the clinical outcomes differ, EHEC and EPEC share a common strategy as they
47
colonize the intestinal gut mucosa via attaching and effacing (A/E) lesions. A/E lesions are
48
characterized by local effacement and destruction of the enterocyte brush-border microvilli,
49
intimate attachment to gut enterocytes and formation of pedestal-like structures beneath
50
attached bacteria through host cytoskeletal rearrangement (3). The ability to cause A/E
51
lesions is associated with the Locus of Enterocyte Effacement (LEE) pathogenicity island
52
which encodes a type 3 secretion system (T3SS), gene regulators, the adhesin intimin,
53
chaperons and 7 effector proteins (4-6). Through T3SS translocation of the effectors, EPEC
54
and EHEC subvert and manipulate the host cell signalling during infection, including
55
intracellular trafficking, apoptosis and inflammation (7).
56
A key regulator of inflammation is the transcriptional regulator nuclear factor-kappa B (NF-
57
ĸB) (8, 9). The major subunits of NF-κB are p65 and p50; in the inactive state p65/p50 are
58
bound to IĸB (inhibitor of κ light chain gene enhancer in B cells) and are restricted to the
59
cytoplasmic compartment (9, 9). The NF-κB signalling pathway is activated in response to a
60
number of stimuli including stress signals, pro-inflammatory cytokines interleukin-1 beta (IL-
61
1β) and tumor necrosis factor (TNF) and pathogen associated recognition patterns (PAMPs)
62
(9). Activation of the NF-ĸB signalling pathways converges on IκB Kinase (IκK) and
63
phosphorylation of IĸBα, which leads to its proteasomal degradation. This results in nuclear
64
translocation of NF-κB (p65/p50) (9) and expression of multiple pro-inflammatory cytokines
65
including interleukin-8 (IL-8) (9).
3
66
The effector EspT, present in the atypical EPEC strain E110019, induces an NF-κB
67
dependent pro-inflammatory response through the activation of the GTPase Rac1 (10).
68
However the pro-inflammatory activity of EspT is restricted to infection of immune cells and
69
furthermore EspT is only present in 1% of clinical EPEC isolates (10). More commonly,
70
multiple recent publications have shown that a number of conserved EPEC and EHEC
71
effectors, particularly NleC and NleE, inhibit TNF and IL-1β-induced NF-κB activation (7).
72
NleC, a zinc metalloprotease, directly cleaves the p65 subunit of NF-κB (11, 12); NleE,
73
through its methyltransferase activity, inhibits transforming growth factor-beta-activated
74
kinase 1 (TAK1) and downstream degradation of IκBα (13-16). In addition, NleH1 targets the
75
human ribosomal protein S3 (RPS3), a co-factor of NF-κB, repressing nuclear translocation
76
of NF-κB (17). Despite inhibition by these effectors some evidence suggests the NF-κB
77
pathway is also activated during EPEC infection in a T3SS-dependent manner. For example
78
infection with the EPEC E2348/69 ΔescN mutant (T3SS deficient) does not promote p65
79
translocation to the nucleus (15). In contrast, the EPEC E2348/69 ΔPP4/IE6 double-island
80
mutant, which lacks the genes encoding the effectors NleE, NleB1, EspL, NleB2, NleC, NleD
81
and NleG, is able to significantly activate NF-κB (12, 15). This suggests that the NF-κB
82
pathway is activated and inhibited in a T3SS-dependent manner; however, no effector has yet
83
been attributed to the activation of this pathway by EPEC during infection of epithelial cells.
84
The non-LEE encoded effector F (NleF) (18, 19) is highly conserved among EPEC and
85
EHEC strains (20). NleF can inhibit the proteolytic activity of the inflammatory caspases;
86
caspase-4 and -8, as well as the activity of caspase-9 (21). The C-terminal 4 amino acids of
87
NleF, with the motif LQCG, are essential for this interaction (21). Furthermore Olsen and
88
colleagues identified TMP21, a type-1 transmembrane protein involved in vesicular transport,
89
as a binding partner of NleF (22). The aim of this study was to further characterize the
90
biological activity of NleF.
4
91
Materials and Methods
92 93
Bacterial strains, growth conditions and reagents
94
The E. coli strains used in this study are listed in Table 1. E. coli were cultured in Luria Broth
95
(LB) at 37⁰C, at 200 rpm with the appropriate antibiotics: 100 µg/ml ampicillin and 50 µg/ml
96
kanamycin. All reagents were purchased from Sigma unless otherwise stated.
97 98
Tissue Culture
99
HeLa cells were grown in Dulbecco's Modified Eagle's Medium - low glucose (1000 mg/L;
100
DMEM) supplemented with 10% (v/v) heat inactivated foetal calf serum (FCS; Gibco), 2mM
101
Glutamax (Invitrogen) and 0.1 mM non-essential amino acids in a 5% CO2 atmosphere.
102 103
Plasmid construction
104
The plasmids and primers used in this study can be found in Table 1 and Table 2,
105
respectively. nleF was amplified from E. coli O127:H6 E2348/69 genomic DNA by PCR
106
using KOD Hot Start Polymerase (Novagen). nleF and nleF1-185 were cloned into the
107
eukaryotic expression vector pRK5::myc following amplification with the primer pair 1F/1R
108
and 1F/2R respectively. PCR products were cloned into the BamHI/PstI sites of pRK5::myc
109
to generate plasmids pICC1661 and pICC1667, respectively. Alternatively nleF was cloned
110
into pEGFP-C2 following amplification with primer pair 7F/7R using the restriction sites
111
EcoRI/BamHI to generate pICC1675. nleF was cloned into the EcoRI/HindIII sites of
112
pBAD22 using the primer pair 6F/6R to construct the plasmid pICC1662. Plasmids were
113
codon optimized to remove a poly-thymine rich region using the primer pair InvF/InvR. All
114
constructs were verified by DNA sequencing.
5
115
Mutant construction
116
nleF was deleted from EPEC E2348/69 using the lambda red system (23), generating strain
117
ICC1082. The primer pair pKD4F/R was used to amplify the Kanamycin cassette from
118
pKD4. Primer pair 3F/R were used to amplify nleF from Escherichia coli O127:H6 E2348/69
119
genomic DNA with a 500 bp 5’ and 400 bp 3’ flanking sequences. The PCR product was
120
cloned into TOPO Blunt II (Invitrogen) by blunt ligation. nleF was removed by inverse PCR
121
with the primer pair 4F/R and replaced with the Kanamycin cassette using blunt end ligation
122
to construct the plasmid pICC1674. The linear product for recombination was amplified from
123
pICC1674 using primer pair 3F/R and mutagenesis was confirmed by sequencing using the
124
primer pair 5F/R.
125 126
Transfection and siRNA
127
HeLa cells were grown until ~40% confluent in 24 well plates containing 13 mm glass cover-
128
slips (VWR International). 16 h post seeding monolayers were transfected with the
129
eukaryotic expression vectors listed in Table 1 using the Gene Juice kit (Novagen/Merck), to
130
the manufacturer’s specifications. Transfected cells were left for 24 h and where indicated
131
stimulated with TNF at 20 ng/ml for 30 min for immunofluorescence or 16 h for RT-qPCR
132
before washing 3 times in PBS. Plates were stored at -80⁰C for RT-qPCR or fixed for
133
immunofluoresence as described below. For knockdown ON-TARGETplus SMARTpool
134
scrambled (D-001810-10-05), caspase-4 (L-004404-00-0005), -8 (L-003466-00-0005) and -9
135
siRNA (L-003309-00-0005; Dharmacon) were transfected at 20 nM using Hi-perfect
136
(Qiagen) following the manufacturer’s instructions. Knockdown efficiency was analysed by
137
RT-qPCR and Western blotting as described below.
138 139
Isolation of mRNA and RT-qPCR
6
140
Cells were washed in PBS and mRNA was isolated following manufacturer’s instructions
141
(Qiagen; RNeasy mini kit). Samples were treated with DNase (Promega) at 37⁰C for 30 min
142
followed by 15 min at 72⁰C. Reverse transcription PCR was carried out following the
143
manufacturer’s instructions (Promega) with the following adaptations: RNasin (Promega)
144
was added to the reaction buffer at 2-4 U/µl. IL-8, caspase-4, caspase-8, caspase-9 and
145
GAPDH cDNA was amplified with primer pairs IL-8F/R, CASP4F/R, CASP8F/R,
146
CASP9F/R and GAPDHF/R, respectively, by RT-qPCR using the 7300 Real Time PCR
147
System (Applied Biosystems) under standard cycle conditions. Changes in gene expression
148
levels were analysed relative to the control (see figure legend for individual experiments)
149
using GADPH as a standard and using the ΔΔCT method.
150 151
Infection assay
152
HeLa cells were washed three times with PBS 3 h prior to infection before replacing the
153
media with DMEM with no additives. Bacterial strains grown overnight for 17 h in LB were
154
diluted 1:100 and primed in DMEM with no additives for 3 h at 37⁰C, in a 5% CO2
155
atmosphere without agitation. Arabinose or IPTG were added at 1 mM or 0.5 mM to ∆nleF
156
pnleF and ∆PP4/IE6 pnleE bacterial cultures 0.5 h before infection, respectively. After 3 h of
157
priming the media was removed from the HeLa cells and replaced with 0.5 ml of a 1:3
158
dilution of primed bacterial culture in DMEM. The infection was carried out for 1.5 h and 4
159
h, and where indicated treated with 20 ng/ml of TNF for 30 mins after 3.5 h of infection. For
160
the 1.5 h or 4 h time point cells were washed 3 times with PBS 1 h post infection and
161
incubated for a further 0.5 h or 3 h. Cells were washed 3 times in PBS and either fixed for
162
immunofluorescence or lysed for Western blot analysis as described below.
163 164
Immunofluorescence staining and antibodies
7
165
For immunofluorescence staining cells were fixed with 3% paraformaldehyde and washed a
166
further 3 times in PBS before quenching in 50 mM NH4Cl, permeabilizing with 0.2% Trition-
167
X100 and washing 3 times in PBS. Coverslips were blocked in 1% Bovine serum albumin
168
(BSA)/PBS. All antibodies, unless stated otherwise, were diluted directly in 1% BSA-PBS.
169
Primary antibodies were incubated for 45 min followed by incubation of the secondary
170
antibodies and reagents for 30 min. Mouse monoclonal antibodies anti-Myc clone 4A6 (05-
171
724; Millipore), anti-caspase-4 clone 4B9 (sc-56056; Santa Cruz), Myc-FITC Clone 9E10
172
(F2047-100), anti-Alpha-Tubulin clone DM1A (T6199), anti-TOMM22 (ab57523; Abcam),
173
rabbit polyclonal antibodies anti-N-terminal NF-κB p65 clone A (sc-109; Santa Cruz), and
174
anti-IκBα (Cell signalling; 2942) were used as primary antibodies. Donkey anti-mouse IgG
175
(H+L)
176
immunofluorescence and Goat anti-mouse-HRP (Jackson Immunoresearch; 115-035-008)
177
and Goat anti-rabbit-HRP for western blot analysis (Jackson Immunoresearch; 111-035-008)
178
were used as secondary antibodies. Phalloidin-TRITC (P1951-.1MG; Sigma) and DAPI were
179
used to visualize actin and the nucleus respectively. Cover-slips were then mounted using
180
Prolong Gold anti-fade reagent (Invitrogen) before visualising on the Zeiss Axioimager
181
immunofluorescence microscope using the Zeiss Axiovision Rel 4.5 software.
conjugated
to
DyLight-488
(715-485-150;
Jackson
Immunoresearch)
for
182 183
Western blotting
184
90 minutes post infection cells were lysed in loading dye (LDS; 250 mM Tris pH 6.8, 40%
185
glycerol, 6 % SDS, 1 % Bromophenol blue and 0.8% β-mecaptoethanol) and boiled for 10
186
mins, run on SDS page and then transferred to a PVDF membrane. The membranes were
187
blocked in Tris buffer saline pH 7.4 (10 mM Tris, 150 mM NaCl)/0.1% Tween (TBS-T)
188
containing 5% semi-skimmed milk (blocking buffer) for 1 h before incubation with primary
189
antibody in blocking buffer for 16 h at 4⁰C. The membrane was washed in TBS-T and the
8
190
secondary antibody was incubated for 2 h in blocking buffer. Western blots were visualized
191
by the addition of ECL solution (GE Healthcare) and developed by the LAS-3000 imager
192
(Fuji). Densitometry was analysed using ImageJ.
193 194
Luciferase assay
195
An NF-κB dual luciferase reporter system (Promega) was employed as described previously
196
(15), with the following adaptations. HeLa cells were seeded as desribed above and co-
197
transfected with pEGFP-C2 or pEGFP-nleF together with pNF-κB-Luc (Clontech) and pRL-
198
TK (Promega). 24 h post transfection, for each condition, the expression of firefly luciferase
199
was measured and normalized for Renilla luciferase measurements and NF-κB activity was
200
expressed relative to HeLa-pEGFP.
201 202
Statistics
203
All data was analysed using GraphPad Prism software, using one-way Annova and
204
Bonferroni’s Multiple comparison test when required. A P