The Cofilin Phosphatase Slingshot Homolog 1 (SSH1) Links NOD1 Signaling to Actin Remodeling
et al. (2014) The Cofilin Phosphatase Slingshot Homolog 1 (SSH1) Links NOD1 Signaling to Actin
Remodeling. PLoS Pathog 10(9): e1004351. doi:10.1371/journal.ppat.1004351
The Cofilin Phosphatase Slingshot Homolog 1 (SSH1) Links NOD1 Signaling to Actin Remodeling
Harald Bielig 0
Katja Lautz 0
Peter R. Braun 0
Maureen Menning 0
Nikolaus Machuy 0
Christine Bru gmann 0
Sandra Barisic 0
Stephan A. Eisler 0
Maria Andree 0
Birte Zurek 0
Hamid Kashkar 0
Philippe J. Sansonetti 0
Angelika Hausser 0
Thomas F. Meyer 0
Thomas A. Kufer 0
Andreas J. Baumler, University of California, Davis, United States of America
0 1 Institute for Medical Microbiology, Immunology and Hygiene , Cologne, Germany , 2 Department of Molecular Biology, Max Planck Institute for Infection Biology , Berlin, Germany, 3 Steinbeis-Innovationszentrum Center for Systems Biomedicine, Falkensee, Germany , 4 Institute of Cell Biology and Immunology, University of Stuttgart , Stuttgart, Germany , 5 Unite de Pathoge nie Microbienne Mole culaire, Institut Pasteur, Paris, France, 6 INSERM U786, Institut Pasteur, Paris, France, 7 Microbiologie et Maladies Infectieuses, Colle`ge de France , Paris , France , 8 University of Hohenheim, Institute of Nutritional Medicine , Stuttgart , Germany
NOD1 is an intracellular pathogen recognition receptor that contributes to anti-bacterial innate immune responses, adaptive immunity and tissue homeostasis. NOD1-induced signaling relies on actin remodeling, however, the details of the connection of NOD1 and the actin cytoskeleton remained elusive. Here, we identified in a druggable-genome wide siRNA screen the cofilin phosphatase SSH1 as a specific and essential component of the NOD1 pathway. We show that depletion of SSH1 impaired pathogen induced NOD1 signaling evident from diminished NF-kB activation and cytokine release. Chemical inhibition of actin polymerization using cytochalasin D rescued the loss of SSH1. We further demonstrate that NOD1 directly interacted with SSH1 at F-actin rich sites. Finally, we show that enhanced cofilin activity is intimately linked to NOD1 signaling. Our data thus provide evidence that NOD1 requires the SSH1/cofilin network for signaling and to detect bacterial induced changes in actin dynamics leading to NF-kB activation and innate immune responses.
Funding: This work was supported by grants of the DFG (SFB670) and the Koeln Fortune Program/Faculty of Medicine, University of Cologne to TAK and the
Heidelberg Academy of Science and Humanities to AH (WIN-Kolleg). The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
. These authors contributed equally to this work.
Effective immune defense in mammals relies on the detection of
conserved pathogen structures by pattern recognition receptors
(PRRs) of the innate immune system to prime immune responses
Several PRRs have been identified and extensively studied in
the last decade. In particular, members of the NOD-like receptor
(NLR)-family gained attention due to their intracellular
localization [2,3]. One of the first NLRs shown to act as a PRR is NOD1.
NOD1 is an intracellular protein that can be activated by
diaminopimelic acid-containing peptides derived from bacterial
peptidoglycan and acts as a sensor for invasive bacteria such as
Shigella flexneri .
A wealth of data suggest that NOD1 is an important PRR for a
variety of bacteria in mammals, which also contributes to systemic
activation of neutrophils, induction of adaptive immunity and
immune tissue homeostasis (reviewed in [2,3]). Upon activation,
NOD1 forms a complex with the receptor-interacting serine/
threonine-protein kinase 2 (RIP2), which results in the activation
of NF-kB and mitogen-activated protein kinases (MAPK) signaling
pathways [2,3]. Several components of the pathway downstream
of NOD1 have been identified. For example, the NOD1 binding
partner RIP2 mediates activation of the TGF-b-associated kinase
1 (TAK1) complex which is induced by ubiquitylation of RIP2
through ubiquitin ligases including the X-linked inhibitor of
apoptosis protein (XIAP)  and the cellular inhibitor of apoptosis
protein-1 and -2 (cIAP1, and cIAP2) . NOD1 is found at the
plasma membrane where it co-localizes with F-actin. This
localization was suggested to be a prerequisite for signaling
because affecting actin polymerization changes NOD1 signaling
. Furthermore, the Salmonella effector SopE activates NOD1,
involving changes in small Rho GTPase activity .
Additionally, the RhoA guanine nucleotide exchange factor H1 (GEF-H1)
was linked to NOD1 activation . Of note, the NOD1 related
protein NOD2 is also regulated by the small GTPase Rac1 [12,13]
and localizes at the plasma membrane at cortical F-actin
structures, similar to NOD1 [9,13,14]. Together this indicates
an intimate connection of NOD1 and NOD2 signaling with the
actin cytoskeleton, although the mechanistic details remain largely
elusive. Cellular actin dynamics are strictly controlled by the
action of nucleation factors such as Arp2/3, which bind to the
sides of pre-existing filaments and promote the growth of new
filaments at these sites. Actin binding proteins belonging to the
actin depolymerization factors (ADF)/cofilin family control the
disassembly of actin filaments by severing F-actin filaments,
thereby generating new sites of actin polymerization. In addition,
there is evidence that cofilin depolymerizes F-actin to provide new
NOD1 was one of the first NLR-family members shown to
act as an important intracellular pattern-recognition
molecule mediating antimicrobial activities in mammals.
It has been demonstrated that perturbation of F-actin and
RhoGTPase activity affects NOD1 and NOD2 signaling,
however, the effectors of this process remained elusive. By
using a multilayered high-throughput druggable genome
wide siRNA screening approach to discover novel
components specific for the NOD1 pathway, we identified the
cofilin phosphatase SSH1, which acts downstream of
RhoA-ROCK, as key regulator of NOD1 signaling. We show
that SSH1 forms a complex with NOD1 at F-actin rich sites
in human cells and is needed for NOD1-mediated
responses towards TriDAP exposure and Shigella flexneri
infection. Functionally this is achieved by SSH1-mediated
activation of cofilin. Our findings reveal a previously
unrecognized role for SSH1 in NOD1 signaling and provide
a plausible unifying mechanistic explanation of how
perturbations of the actin cytoskeleton can induce
NOD1-mediated inflammatory responses.
G-actin molecules for polymerization. Cofilin activity itself is
tightly controlled by LIMK1 and LIMK2, which phosphorylate
cofilin at serine 3 whereby its activity is blocked. Accordingly,
dephosphorylation by the phosphatase slingshot homolog 1
(SSH1) reactivates cofilin (reviewed in ).
Here we identify the cofilin phosphatase SSH1 as an essential
component of the human NOD1 signaling pathway and show that
SSH1 links NOD1 activation to cofilin-mediated changes in actin
A high-throughput siRNA-screen identifies SSH1 as an
essential component of NOD1-mediated NF-kB signaling
To identify novel factors involved in NOD1-mediated NF-kB
activation, we adapted a cell based NF-kB-luciferase reporter gene
assay in HEK293T cells  for high-throughput (HT) small
interfering RNA (siRNA) screening (Figures S1A and B).
A druggable-genome siRNA-library (a sub-library of the human
genome covering approximately 7000 genes with known protein
domains) containing four independent siRNAs per gene was
screened in quadruplicate for hits inhibiting NOD1-mediated
NFkB activation upon treatment with the NOD1-specific elicitor
TriDAP (Figure S1A). After quality control and elimination of
toxic siRNAs, preliminary candidates were selected using a
probability-based algorithm (redundant siRNA activity; RSA)
 (Table S1). Statistical analyses confirmed the high
reproducibility of the results and the robustness of the assay controls (p65
and the non-targeting Allstars siRNA) (Figure S2A). The top 435
of the RSA-ranked candidates with at least two hit siRNAs were
differentially tested for TriDAP- as well as TNF-induced NF-kB
activation (referred to here as validation screen and counter
screen, respectively), using two independent siRNAs in
HEK293T cells (Figure 1A and S1B). Knock-down of 173 genes
for this set displayed an inhibitory effect on TriDAP-induced
NFkB activation. Of those, 66 genes were specifically involved in
NOD1 signaling in HEK293T cells, i.e. they did not significantly
affect TNF-mediated NF-kB activation (Figure 1B, Table S1).
Among them, 28 known regulators of NF-kB, partly with known
specificity for NOD1 signaling, could be retrieved, confirming the
validity of the screening results (Table S1). Gene ontology (GO)
enrichment analysis revealed that GO terms linked to immune
function, and in particular to NLR function, were significantly
overrepresented among the preliminary hits (Figure S2B).
Indepth analysis using Ingenuity pathway analysis highlighted that
56 of the 435 preliminary screen hits (12.9%) are known
components of NF-kB signaling (Figure S2C). Among these 56
preliminary hits, 28 could be validated in the HEK293T
validation screen, of which 14 were not influencing
TNF-ainduced NF-kB activation.
In order to further minimize false positive hits due to off-target
effects of the used siRNAs and also to exclude a cell type specific
bias, the 435 preliminary candidates were further tested for their
effect on endogenous NOD1-mediated NF-kB activation in
human myeloid THP1 cells (THP1-blue reporter line) (Figure
S2D), revealing a cluster of genes showing functional interactions
as revealed by STRING analysis (Figure S2E). The results
confirmed 28 genes that showed an effect on NOD1-mediated
NF-kB activation in both HEK293T and THP1 cells. Among
those, receptor-interacting serine/threonine-protein kinase 2
(RIPK2), NOD1, transcription factor p65 (RELA), X-linked
inhibitor of apoptosis protein (XIAP), deltex 4 (DTX4), calreticulin
(CALR), olfactory receptor family 12, subfamily D, member 2
(OR12D2), and ring finger protein 31 (RNF31) were the strongest
candidates (.3 S.D.). Exclusion of the genes that affected
TNFinduced NF-kB activation in HEK293T cells resulted in a short list
of 18 genes (Figure 1C, Table S1). Although this candidate-list
obtained in THP1 cells differed from that derived from HEK293T
cells, the genes RIPK2, XIAP, the uracil nucleotide/cysteinyl
leukotriene receptor (GPR17), SSH1, the snail family zinc finger 1
(SNAI1), and CHUK (IKKa) were confirmed as hits in both cell
lines by this highly stringent procedure (Figure 1C,Table S1).
Identification of RIPK2, XIAP and the inhibitor of nuclear factor
kappa-B kinase subunit alpha (IKKa), all of which have recently
been linked to the NOD1 pathway [7,18,19] validated the success
of the screening procedure.
To demonstrate that we can reversely reproduce the findings
from the screen, XIAP was silenced with a screen-independent
siRNA. This strongly impaired NF-kB activation upon both
NOD1 or NOD2 stimulation in HEK293T cells (Figure S3A and
B) and THP1-blue cells (Figure S3C). Furthermore, TriDAP- or
MDP-induced IL-8 secretion in THP1 cells was significantly
reduced (Figure S3D). Taken together, our screen validated XIAP
as an essential component of the NOD1 signaling pathway, in line
with previous reports showing an involvement of XIAP in NOD1
and NOD2 signaling [7,18]. XIAP was shown to mediate linear
ubiquitylation of RIP2  mediated by the so-called LUBAC
complex. Of note, our screen also highlighted one component of
the LUBAC complex, RNF31 (HOIP) as a strong candidate
(Table S1), which was independently identified in a recent screen
for components of the NOD2 signaling pathway . Notably, we
identified three novel genes, namely the uracil nucleotide/cysteinyl
leukotriene receptor GPR17, the cofilin phosphatase SSH1 and
the transcriptional repressor SNAI1 that hitherto had not been
reported in the context of NOD1 signaling.
SSH1 specifically contributes to NOD1- and
NOD2mediated inflammatory responses
Among these stringently validated hits, the phosphatase SSH1
(Table S1), a key regulator of actin dynamics (reviewed in )
caught our attention. In order to validate SSH1 as a critical
component of NOD1 signaling, we depleted the protein by two
different siRNAs in myeloid-like differentiated THP1-blue cells
and measured NF-kB activity and IL-8 release upon treatment
with the NOD1, NOD2, TLR4, and TNFR agonists TriDAP,
Figure 1. HTS siRNA-screening identifies SSH1 as novel component of the NOD1-pathway. (A) Combined Z-scores of the primary-hits in
the counter screen with TriDAP (NOD1,ordinate) and TNF (abscissa) activation. (B) Flow-chart representing the screening procedure. Number of hits
(genes) of each step is indicated. (C) Final hitlist ranked according to the THP1 results (best). Candidates validated in all steps and not affecting TNF
signaling are shown in bold. Z-score were normalized to control siRNAs set to 0. (see also Figures S1, S2 and Table S1).
MDP, LPS, and TNF, respectively (Figure 2A and B). This
revealed a significant reduction of NOD1- and NOD2-mediated
inflammatory responses (Figure 2A and B) concurrent with a
reduction in SSH1 levels by siRNA treatment (Figure 2C). In
contrast, TNF- and TLR4-induced responses were not highly
significantly affected by reduced SSH1 levels (Figure 2A and B).
Similar results were obtained in THP1 cells not containing the
reporter construct, showing that NOD1-mediated release of
several key inflammatory cytokines was reduced upon SSH1
depletion (Figure S4).
Immunoblotting showed that both siRNAs decreased the
expression of SSH1 and revealed that SSH1 protein levels were
increased upon PAMP stimulation (Figure 2C). To evaluate the
effect of SSH1 knock-down at early time points after activation, we
measured IL-8 and SSH1 mRNA levels 3 h after TriDAP or TNF
stimulation in THP1-blue cells. This showed that the reduction in
SSH1 correlated with reduced IL-8 mRNA levels when cells were
activated by TriDAP but not when stimulated with TNF
To elucidate the effect of SSH1 on physiological activation of
NOD1 by bacteria, we infected HeLa cells with the
Gramnegative invasive bacterial pathogen Shigella flexneri, which is
sensed primarily by NOD1 in these cells. Depletion of SSH1
mRNA by the two SSH1-specific siRNA duplexes resulted in
significantly impaired IL-8 and IL-6 production 6 h post infection
with the invasive S. flexneri strain M90T (Figure 3A). Notably,
this was not due to reduced bacterial invasion and replication, as
demonstrated by gentamicin protection assays (Figure 3B). To
decipher if SSH1 is involved preferentially in early or late events
during the inflammatory response to S. flexneri, we measured
IL8 mRNA and IL-8 release at different time points after bacterial
infection. Depletion of SSH1 (Figure S5A) led to greatly reduced
IL-8 transcription as early as 30 min after infection (Figure 3C).
Two hours post infection, when IL-8 became measurable in the
supernatant of infected cells, a significant difference in IL-8
secretion was observed between cells treated with SSH1 targeting
siRNA and those treated with a non-targeting control siRNA
(Figure 3C). This suggests that SSH1 is already involved at early
Figure 2. SSH1 is essential and specific for NOD1-mediated signaling. (AB) PMA differentiated THP1-blue cells were treated for 72 h with a
non-targeting (siCTRL) or two SSH1-specific siRNA duplex and incubated with TriDAP (10 mg/ml), MDP (10 mg/ml), TNF (0.1 mg/ml), or LPS (0.05 mg/
ml). (A) NF-kB activation was measured by SEAP secretion. (B) IL-8 levels in the culture supernatants of the cells from (A). Values are mean + S.D. from
two independent experiments conducted in triplicates. Significance was calculated by students t-test (unpaired, two-tailed) *p,0.05, **p,0.005. n.s.:
not significant. (C) Immunoblot of one of the experiments from (A), probing for SSH1 and actin as loading control is shown. (D) Early effects of the
knock-down of SSH1. IL-8 (left panel) and SSH1 (right panel) mRNA levels in THP1-blue cells treated as inducted were measured by qPCR. Mean + S.D.
from triplicate measurements of one representative experiment is shown (see also Figure S4).
times of NOD1 activation. As expected, the knock-down of SSH1
led to a strong increase in phosphorylation of its substrate cofilin at
serine 3 (Figure S5B), proving that SSH1 is active in our cellular
system. To rule out that the effects described above were related to
the malignant origin of the cell lines, we assessed the phenotype of
SSH1 depletion in primary human dermal fibroblasts. In line with
the results obtained with cell lines, in primary human dermal
fibroblasts NOD1-mediated IL-8 secretion was also significantly
reduced upon SSH1 knock-down, whereas SSH1 depletion
affected TNF-induced IL-8 secretion to a lesser extent (Figure
Taken together, our data identified SSH1 as an essential
component of NOD1-mediated activation of pro-inflammatory
responses in human myeloid and epithelial cells.
SSH1 interacts with NOD1 at F-actin structures
SSH1 was reported to bind F-actin and to co-localize with actin
stress fibres [22,23]. We confirmed that ectopically expressed
SSH1 partly co-localized with F-actin in HeLa cells that stably
express YFP-NOD1. Although the two proteins showed slightly
different localization patterns, a partial co-localization of SSH1
and NOD1 was also observed in confocal imaging (Figure S6A). In
co-immunoprecipitation experiments of transiently expressed
SSH1 and NOD1/2 proteins, both NOD1 and NOD2
coprecipitated with SSH1 (Figure 4A), validating that both NLR
proteins are able to form a complex with SSH1. Upon activation
of NOD1 by TriDAP we observed a change in the stoichiometry
of this complex with an increase in affinity about 30 min after
TriDAP stimulation and reduced binding at later time points
(Figure 4B). To analyse this protein-protein interaction in more
detail we employed in situ proximity ligation assays (PLA). We
found that GFP-SSH1 and Flag-NOD1 associated predominantly
at F-actin positive structures (Figure 4C). Quantitative analysis
using high throughput microscopy revealed that the area covered
by PLA spots was approximately 4-fold higher in cells expressing
GFP-SSH1 and Flag-NOD1 than in neighbouring GFP-negative
cells, confirming the specificity of the signal (Figures 4C and S6B).
Moreover, the phosphatase-dead mutant C393S of SSH1 gave
similar results in the PLA assays as WT SSH1 (Figure S6C)
showing that association of SSH1 and NOD1 was not dependent
on the phosphatase activity of SSH1. Accordingly, SSH1 was not
involved in membrane targeting of NOD1 and NOD2 in HeLa
cells, as the sub-cellular localization of transiently transfected
NOD1 and NOD2 were not markedly changed in cells with highly
reduced SSH1 expression compared to controls (Figure S7A and
It was reported that in epithelial cells SSH1 is enriched at the
entry foci of Salmonella . Likewise, NOD1 is enriched at entry
sites of S. flexneri . We thus analyzed the localization of SSH1
in S. flexneri-infected cells, revealing an enrichment of SSH1 at
the bacterial entry foci (Figure 4D).
We previously reported that depolymerization of F-actin by the
mycotoxin cytochalasin D enhances NOD1-mediated NF-kB
activation . Based on our results, we hypothesized that the
impact of SSH1 on NOD1 signaling might be mediated by
cofilincontrolled actin remodeling and that the formation of a
SSH1NOD1/2 complex might control the local NOD1/2 activity at
sites of bacterial entry. To address this, we monitored SSH1
activity indirectly by measuring phosphorylation of its substrate
cofilin at serine 3 (p-cofilin) after TriDAP-induced activation of
NOD1. TriDAP treatment of HEK293T cells expressing low
levels of NOD1 strongly induced IL-8 release from these cells
(Figure 4E). Interestingly, basal p-cofilin levels decreased at the
time when IL-8 transcription was highest (about 180 min post
TriDAP treatment) (Figure 4E). After longer periods of incubation
(16 h), p-cofilin levels increased above the level seen in untreated
cells (Figure 4E). The slower kinetics in these experiments
compared to S. flexneri infection in HeLa cells (Figure 3) is likely
due to the different kinetics of TriDAP uptake in HEK293T cells
. Notably, no obvious change in p-cofilin levels was observed
after stimulation of HEK293T cells with TNF, although TNF
induced high IL-8 release (Figure 4F and S6D). Moreover, in cells
in which NOD1 was depleted by siRNA TriDAP treatment did
not robustly influence p-cofilin levels over time (Figure 4F).
Accordingly, TriDAP failed to induce IL-8 secretion form these
cells (Figure S6D). Moreover, expression of SSH1 did not result in
increased basal NF-kB activity (Figure S8B).
Collectively, these data provide evidence that NOD1 signaling
relies on the presence of SSH1 that acts downstream of NOD1
and involves the activation of cofilin.
NOD1 signaling is connected to changes in the actin
We next asked if other components of the cofilin regulatory
network also contributed to NOD1 signaling outcome. Using
siRNA-mediated knock-down we confirmed that reduction of
cofilin resulted in a similar perturbation of NOD1 signaling as
SSH1 knock-down (Figure 5A). Cofilin is regulated primarily by
phosphorylation through the kinases LIMK1/2, which is
counteracted by the phosphatase activity of SSH1. Both SSH1 and
LIMK1/2 are themselves regulated by phosphorylation events
mediated by protein kinase D (PKD), ROCK1/2 and PAK1/4,
respectively (reviewed in ). Expression of a dominant negative
form of ROCK1 (KD-IA, which lacks kinase and Rho binding
activity) enhanced, albeit not significantly, NOD1-mediated
responses in a dose-dependent manner, whereas a constitutively
active mutant of ROCK1 (delta1) significantly inhibited signaling
by NOD1 (Figure 5B). To further substantiate this, we tested the
effect of two potent chemical inhibitors of ROCK - Y-27632 and
Glycyl-H1152 - on NOD1-mediated signaling. In HEK293T cells,
both inhibitors enhanced TriDAP-induced NOD1-mediated
NFkB activation in a dose dependent manner, this enhancement was
significant in the case of Glycyl-H1152 (Figure 5C). By contrast,
both compounds led to a significant reduction of TNF-induced
NF-kB responses (Figure S8A). Chemical inhibition of ROCK also
led to higher NOD1-mediated pro-inflammatory responses in
TriDAP stimulated HeLa and THP1 cells (Figures 5D and E).
ROCK inhibition correlated in a dose dependent manner with
reduced levels of p-cofilin, suggesting that increased cofilin activity
causes this effect on NOD1 signaling. As shown before, stimulation
of cells with TriDAP further enhanced dephosphorylation of
cofilin (Figures 5D and E). To provide direct evidence that the
phosphatase activity of SSH1 is responsible for modulation of
NOD1 activity, we overexpressed the phosphatase-dead mutant
C535S of SSH1 in HEK293T cells. In line with our hypothesis,
this did not affect NOD1-mediated signaling (Figure S8C).
These results show that perturbation of the cofilin pathway at
different levels affected NOD1 signaling, suggesting that NOD1
signaling relies on cofilin-mediated changes in actin remodeling.
Accordingly, NOD1 signaling induced by actin
polymerizationperturbing mycotoxins should be SSH1 independent. As reported
for HEK293T cells , we found that depolymerization of F-actin
using cytochalasin D strongly enhanced NOD1-mediated
signaling in THP1 cells (Figure 6A). In line with previous reports,
cytochalasin D enhanced IL-8 release induced by other PAMPs
about 2-fold [26,27]. However, in the case of NOD1 activation by
TriDAP, a significantly higher increase to ,3-fold was observed
(Figure 6B). Notably, this was not the case upon activation of
NOD2 by MDP. Next, we depleted SSH1 expression in
THP1blue cells and subsequently disturbed actin polymerization by
cytochalasin D treatment. Knockdown of SSH1 significantly
reduced NOD1-mediated NF-kB activity in cells treated with
TriDAP, however, treatment with cytochalasin D rescued the
effect of SSH1 depletion on TriDAP-induced NOD1 activation
(Figure 6C). This strongly suggests that F-actin affects NOD1
signaling downstream of SSH1.
Taken together, our data support that NOD1 activation and
induction of pro-inflammatory responses requires actin
remodeling controlled by the SSH1 and cofilin network.
Using an unbiased high-throughput siRNA screen, we identified
novel factors involved in the regulation of the NOD1 signaling
cascade. The validity and quality of our screening approach is
highlighted by the fact that the screen identified many factors
involved in canonical NF-kB and/or NOD1 signaling. Most
prominently, the primary screen validated the proteins RIPK2
(reviewed in ), IKKa , IKKb (reviewed in ), TAB2
[28,29], RNF31 , p50 and RELA  as essential positive
regulators of NOD1 signaling. Beside RIP2, XIAP ranked the
highest throughout the whole screening procedure among the
NOD1 specific hits. In line with a recent publication, we observed
blunted responses of XIAP depleted cells to stimulation with
NOD1 and NOD2 elicitors . A recent study now provides the
framework for the function of XIAP in this process, by showing
that it acts as a ubiquitin ligase for RIP2, catalyzing linear
ubiquitylation events, at least in NOD2 signaling . By using S.
flexneri as an infection model we could recently demonstrate the
physiologic relevance of these findings in vitro and in vivo,
underscoring the impact of XIAP on anti-bacterial immunity .
With high confidence the screen identified a central regulator of
actin cytoskeletal dynamics, the phosphatase SSH1, as novel
analysis using a specific antibody (lower panel). Detection of GAPDH served as loading control. (E) Same experimental setting as (D), using HeLa cells
stimulated with 10 mg/ml TriDAP. Mean + S.D. is shown. * p,0.05, ** p.0.005 compared to Ctrl (see also Figure S8).
component of the NOD1 signaling cascade. Notably, SSH1 was
recently also identified as a potential hit in two independent
siRNA screening efforts searching for NOD1 and NOD2 signaling
components, although it was not validated in neither of these
studies [20,33]. SSH1s best described function is the
dephosphorylation and subsequent activation of the actin
depolymerization factor cofilin. Its activity is known to be counter-acted by
LIMK1 and LIMK2, which phosphorylate and thus inactivate
cofilin on serine 3 (reviewed in ). Using a highly stringent and
unbiased multilayer screening approach offers high confidence in
the obtained data. However, candidates can be overlooked due to
loss because of mismatch of quality criteria. This likely explains
why cofilin, ROCK1, ROCK2 and others, although being
represented in the screened library, were not identified as
We confirmed the function of SSH1 in NOD1 signaling by
independent siRNA knock-down experiments in different human
cell lines and primary human dermal fibroblasts. This validated
that silencing of SSH1 significantly impaired NOD1-mediated
responses in human cells triggered by TriDAP and infection with
the invasive bacterial pathogen S. flexneri. Consistent with the
screen data, SSH1 knock-down affected TNF and LPS-induced
NF-kB activation to a far lesser extent than NOD1- and
NOD2mediated responses, showing that SSH1 contributes to NOD1 and
NOD2 signaling in a rather specific manner. SSH1 thus
contributed to NOD1 signaling at early times. Taken together,
our results suggest that SSH1 affects NOD1 signaling through its
phosphatase activity. This is best evidenced by the observation that
NOD1-induced activation of IL-8 transcription was accompanied
by a reduction of cofilin phosphorylation and the lack of effect of a
phosphatase-dead mutant of SSH1 on NOD1 signaling. Our data
do not allow drawing conclusions on how SSH1 activity is
triggered in this process. However, the observed complex
formation of SSH1 and NOD1, that exhibited changed
stochiometry upon triggering of NOD1 by its elicitor TriDAP, makes it
tempting to speculate that binding of NOD1 to a SSH1-containing
complex initiates local SSH1 activation. Noteworthy, we observed
that treatment with several PAMPs resulted in enhanced SSH1
proteins levels in THP1 cells. A plausible interpretation of this
finding might be that higher SSH1 levels might render host cells
more prone for enhanced and more rapid NOD1-mediated
immune signaling. Further research will help to address this and to
establish the biological significance of this finding.
Invasive bacteria, such as Shigella and Salmonella, depend on a
tightly controlled spatial and local reorganization of F-actin at the
plasma membrane to gain entry into the host cell. NOD1 is
wellrecognized as an important sensor of bacterial invasion and we
showed earlier that it co-localizes with F-actin at the cell
membrane and that depolymerization of F-actin by cytochalasin
D augments NOD1 signaling in epithelial cells . In the cell,
actin dynamics are controlled by a balance between the activities
of the small GTPases RhoA and Rac1 and it has been reported
that changing their activity affects NOD1 and NOD2 signaling
[12,13,34]. The pathogenic bacterium Klebsiella pneumonia seems
to inhibit Rac1 activity to trigger NOD1 signaling, resulting in a
dampened innate immunity response . Additionally, the Rho
activator guanine nucleotide exchange factor H1 (GEF-H1) was
identified as an essential component of NOD1-mediated signaling
in response to Shigella and muropeptides . In all these studies,
the mechanistic link to the modulation of NOD1 signaling,
however, was not conclusively identified. Our results show that
NOD1 signaling competence relies on actin remodeling via cofilin.
Activation of NOD1 by chemical ligands reduced cellular p-cofilin
at the time when pro-inflammatory signaling was induced.
Because SSH1 and NOD1 directly interact at the plasma
membrane, SSH1 might act as a local platform to recruit
NOD1 to the entry site of pathogens. The fact that cofilin activity
is also modulated by RhoA and Rac1 activity  strongly
suggests that SSH1 and cofilin are key effectors that link NOD1
activation to perturbations in the network of actin regulation. In
support of this notion, our data show that sterile interference
with the cofilin pathway at several levels, as well as
pharmacological disruption of the actin cytoskeleton, modulated NOD1
signaling outcome. For example, we observed enhancement of
NOD1 signaling upon overexpression of a dominant negative
protein of the RhoA effector kinase ROCK or pharmacological
inhibition of ROCK kinases. We cannot formally exclude that
the inhibitors affected other cellular targets. However, in
conclusion all data strongly support that interfering with the
Factin network downstream of RhoA and upstream of cofilin
profoundly alters NOD1 signaling. Finally, we observed that
induction of NOD1-mediated responses by depolymerization of
F-actin is independent of SSH1. Taken together, these
experiments showed that NOD1 signaling outcome correlated directly
with cofilin activity.
F-actin depolymerization by cytochalasin D in myeloid cells was
shown to affect NF-kB signaling in a broader manner, as
confirmed by our results [26,27]. It should be noted, that a
comparative analysis including multiple PRRs was not conducted
in these studies. We observed a much higher synergy on the
NOD1-induced NF-kB activation, indicating that NOD1 signaling
is particularly prone to changes in actin dynamics. Surprisingly,
also NOD2-induced IL-8 responses were less strongly enhanced by
cytochalasin D in myeloid cells compared to NOD1, although
SSH1 interacted with NOD2 and knock-down of SSH1 also
affected NOD2-mediated signaling. Further research is needed to
define the surprising differences in the contribution of the actin
cytoskeleton and SSH1 to NOD1 versus NOD2 signaling.
Regulation of PRR signaling by actin is not without precedence,
as a role for Rac1 in regulation TLR2 function has been shown
before . Furthermore, there are interesting parallels in the
regulation of mammalian and plant NLRs, suggesting that effector
triggered immunity (ETI) in plants brought about by activation of
plant NLR proteins, is also intimately linked to actin dynamics. In
Arabidopsis there is genetic evidence that the actin remodeling
protein ADF-4 negatively affects RPS4-mediated ETI responses,
although the authors do not disclose if this is linked to changed
actin dynamics .
Very recently it has been proposed that NOD1 acts as a
sensor of Rac1 and CDC42 activity induced by bacterial type
III effector proteins, such as the Salmonella virulence factor
SopE . It is, however, still elusive how this is mechanistically
linked to changes in NOD1 activity. In any case,
bacterialinduced perturbation of actin dynamics that are needed for
bacterial cell entry, in particular in epithelial cells, would result
in enhanced NOD1-mediated inflammatory responses. The
data reported here provide novel insights into the underlying
mechanisms showing that SSH1-mediated actin remodeling is a
central component of NOD1 activation and innate immune
Figure 6. NOD1 signaling is dependent on F-actin. (A) Effect of 6 h cytochalasin D (cytoD) treatment on TriDAP/NOD1 induced NF-kB
activation in PMA differentiated THP1-blue reporter cells. Mean + S.D. (n = 3). Significance was calculated by students t-test (unpaired, two-tailed)
*p,0.05, **p,0.005 (compared to 0 nM cytoD). (B) IL-8 secretion of THP1 cells treated for 6 h with the indicated PAMP in the presence or absence of
cytochalasin D. Shown is the mean of triplicates from two independent experiments + S.D. (n = 6). The fold induction of IL-8 by cytoD is depicted
below. (C) SSH1 knock-down effect on cytoD enhanced NOD1 signaling. Differentiated THP1 cells were treated with cytoD for 6 h and stimulated
with 10 mg/ml TriDAP (black bars) or mock treated (open bars). IL-8 secretion measured in the supernatant after 6 h of incubation is shown, mean +
S.D. (n = 3), * p,0.05, ** p,0.005.
Materials and Methods
Cells and bacteria
HEK293T, HeLa, THP1 and THP1-blue (InvivoGen, France)
cells were cultured as described in in . For
immunofluorescence, a HeLa line stably expressing EGFP-tagged NOD1 was
generated. All cell lines were continuously tested for absence of
mycoplasma contamination by PCR. Primary human dermal
fibroblasts were obtained as previously described .
Plasmids and reagents
Plasmids encoding myc tagged human NOD1 and NOD1 were
generated by PCR cloning in a pCDNA3.1 backbone.
SSH1 encoding plasmids are described in . Plasmids
encoding ROCK and mutants are described in  and RhoA
and Rac1 plasmids were a kind gift from Monilola Olayioye
(University of Stuttgart). ROCK inhibitors and cytochalasin D
were purchased from Tocris.
siRNA-based knock-down in HeLa and THP1 cells was
performed as described previously in . siRNAs used:
SSH1_1 : SI00123585, SSH1_3: SI00123599 (Qiagen) and
AllStars negative control (Qiagen).
Bacterial infection with Shigella flexneri
For the infection with S. flexneri, HeLa cells were seeded in
24well plates. Infection was performed using the strain M90T afaE as
described previously . Gentamycin (100 mg/ml) was added to
the cells 30 min after addition of the bacteria. As control, a
noninvasive derivative (BS176 afaE) was used.
Uptake of S. flexneri was analyzed by lysis of infected cells in
0.5% SDS/H2O. Serial dilutions of cell lysates were plated onto
trypticase soy broth bacto agar plates without antibiotics and
incubated at 37uC for 24 h. Colonies were counted and the
recovery was determined.
Co-immunoprecipiation and immunoblot analysis
Immunoprecipitations and immunoblots were conducted using
9E10-agarose (Santa Cruz) essentially as described previously .
Cells were transiently transfected with SSH1 constructs and
NOD1 or NOD2 expression plasmids for 24 h. Antibodies used:
mouse anti-Flag (Sigma, M2), mouse anti-myc (Santa Cruz, 9E10),
HRP-conjugated goat anti-mouse IgG (Bio-Rad) and
HRPconjugated goat anti-rabbit IgG (Bio-Rad), rabbit anti-SSH1
(Abcam, 76943), rabbit anti-cofilin P-S3 (Cell Signaling, 3313),
rabbit anti-cofilin (Cell Signaling, 5175), rabbit beta-actin
HRPconjugated (Santa Cruz, sc-47778 HRP), rabbit anti-alpha tubulin
(Sigma, T7816), rabbit anti-GAPDH (Santa Cruz, 25778).
Luciferase reporter assays
Activation of inflammatory pathways was measured using a
luciferase reporter assay described previously . The means and
standard deviations were calculated from triplicates.
RNA preparation and RT-PCR
Total RNA was extracted from cells using the RNeasy kit
(Qiagen). One mg of RNA was reverse transcribed using the
FirstStrand cDNA synthesis kit (Fermentas).
For quantitative PCR analyses, 50 ng cDNA was analyzed in a
total volume of 25 ml using the iQ SYBR Green Supermix
(BioRad), according to the manufacturers protocol. All quantitative
PCR reactions were run on a Bio-Rad iQ5 cycler, and data were
evaluated by the iQ5 system software (version 2.0) using the
For quantification of SSH1, the following primers were used
(5939): CGTTGCGAAGACAGAATCAA and
CTCCACAGTCGGAGAACCAT. Primer for IL-8, NOD1, NOD2, and GAPDH
are described in .
Indirect immunofluorescence microscopy
Cells were transfected using Lipofectamine 2000 (Invitrogen).
After 1624 h incubation, cells were fixed and processed as
described previously .
DNA was stained using DAPI. Images were acquired on an
Olympus Fluoview 1000 confocal microscope and processed using
In situ proximity ligation assay
The proximity ligation assay was performed with the Duolink
system (Olink bioscience). HeLa cells grown on collagen-coated
coverslips were transfected and fixed 24 h post transfection in 4%
paraformaldehyde for 15 min, washed, permeabilized with 0.1%
Triton X-100, and blocked with blocking buffer (Olink bioscience)
for 30 min at 37uC. The cells were incubated with the primary
antibodies (rabbit GFP-specific antibody and mouse Flag-specific
antibody) diluted in blocking buffer for 2 h. As a negative control,
GFP-SSH1 and Flag-NOD1 transfected cells, which were
incubated in antibody diluent without primary antibodies, were
used. Incubation with secondary antibodies and ligation and
amplification were done as recommended by the manufacturer.
Cells were stained with HCS Cellmask deep red (Life
Technologies) and mounted in mounting medium (Olink bioscience). PLA
dots were acquired with a LSM710 confocal microscope (Zeiss).
FActin labeling (Alexa633-coupled phalloidin, Life Technologies)
was performed after PLA staining. Quantifications of PLA dots
were performed with HCS/HTS based automated PLA spot
detection. In brief, images were acquired on a WiSCAN Hermes
system (Idea Biomedical, Israel), equipped with a Olympus 206
0.75 NA objective. The quantitative spot analysis was performed
using the WiSOFT image analysis software (Idea Biomedical,
Israel). Cells were segmented via the HCS cell mask deep red
staining. After segmentation, cells were classified into
GFPpositive- and negative cells using a threshold in the green channel.
Spots were identified in the red channel and the total area covered
by all spots in one cell was calculated. Spots outside of the cell
mask were not considered.
Detection of cytokines by ELISA
Measurement of cytokines was performed using the appropriate
ELISA kits (Duoset, R&D) according to the manufacturers
instructions. Multiplex cytokine analysis was performed by flow
cytometry using the human inflammation 2plex kit (eBioscience).
Large scale siRNA NF-kB luciferase screen
The siRNA screen was performed using the human
druggable-genome siRNA library from Qiagen (Hilden,
Germany) consisting of four individual siRNAs for each gene.
HEK293T cells were transfected with siRNA (20 nM) using
HiPerFect (Qiagen) and treated with TriDAP (0.5 mM,
InvivoGen, San Diego, CA, USA) or TNF (5 ng/ml), respectively, in
case of the counterscreen.
For screening, cells with a passage number of 2 were used. The
whole assay procedure was performed automatically using a
Biomek FXP laboratory automation workstation (Beckman
Coulter) in 384 well plates (Corning).
4 ml of 200 nM siRNAs (Qiagen) were pre-spotted on 384 well
plates to allow reverse transfection of cells. All plates contained
non-targeting (Allstars; Qiagen) as well as p65, NOD1, and PLK
siRNAs as internal controls. For transfection, 8 ml medium per
well were mixed with 0.25 ml HiPerFect and added to the siRNAs.
The mixtures were incubated at RT for 15 min before adding
1,000 HEK293T cells in 30 ml medium.
After 48 h incubation the medium was changes and cells were
transfected with the NF-kB-luciferase reporter system (11.6 ng
bgal plasmid, 7.03 ml NF-kB-luciferase plasmid, 0.135 ng
NOD1expression plasmid, 8.78 ng pcDNA-plasmid, and 0.0918 ml
Fugene6 (Roche)). Subsequently, cells were stimulated with
0.5 mM TriDAP (InvivoGen) in a volume of 3 ml H2O. For the
TNF-counterscreen, 5 ng/ml TNF were added instead. After
stimulation, cells were incubated at 37uC and 5% CO2 for 16 h.
For read-out, cells were lysed by adding 30 ml 2xlysis buffer
(50 mM Tris pH 8.0, 16 mM MgCl2, 2% Triton, 30% Glycerol,
H2O) and subsequently mixed by pipetting. 35 ml of the lysate
were then added to a white 384 well plate containing 35 ml
reading-buffer (16 lysis buffer w/o Triton containing 0.77 mg/ml
D-luciferin and 1.33 mM ATP). Subsequently, bioluminescence of
the samples was measured with an Envision plate reader
For b-gal read-out, 35 ml of ONPG-development buffer (4 mg/
ml ONPG in 60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM
KCl, 1 mM MgSO4, pH 7.0) were added to the remaining lysate.
After 15 min incubation at 37uC and 5% CO2, the absorbance of
the samples at 405 nm was measured automatically with an
Envision plate reader (PerkinElmer). Each individual siRNA was
tested in 4 four biological replicates.
THP1-blue siRNA screen
For screening, the siRNAs were pre-spotted on 384 well cell
culture plates. All plates contained Allstars, p65, NOD1, and PLK
siRNAs as internal controls. For transfection, 0.25 ml HiPerFect
were added to the siRNAs. The mixtures were incubated at RT
for 15 min, before THP1-blue cells and 0.1 mM PMA (Sigma,
Munich, Germany) were added. The cells were incubated for
72 h, while the growth medium was exchanged twice a day.
Subsequently, cells were stimulated with TriDAP (10 mg/ml,
InvivoGen). The cells were incubated for 16 h before read-out.
For SEAP-detection, 10 ml supernatant per well were transferred
to a plate containing 50 ml QUANTI-Blue SEAP detection
medium (InvivoGen) and incubated for 5 h. 10 ml XTT reagent
were added to the remaining 10 ml and incubated at 37uC for 1 h.
Absorption at 632 nm and 485 nm respectively were measured
using a PerkinElmer EnVision plate reader.
Data analysis of siRNA screen
Data was processed using the CellHTS2 package ,
Bioconductor/R, and Excel (Microsoft). By dividing the luciferase
signal (relative light units; RLU) by the b-gal signal (ABS405),
normalized relative light units (RLU/ABS405 = nRLU) were
achieved. To exclude experimental artifacts, all data from a given
plate was excluded, if the average b-gal signal of the non-targeting
controls (ABS405) was .2.5, ,0.2, or had a standard deviation of
.50%. Next, all wells showing a b-gal signal of ,40% of the
noncoding controls, supposedly due to low plasmid transfection
efficiency or siRNA toxicity, were excluded from further analysis.
Subsequently the nRLU where normalized relative to the
inhibitory effect of the p65-control-siRNAs compared to the
non-targeting controls (normalized percent inhibition; NPI) and
median z-scores of the 4 biological replicates were calculated
(centered to the median of non-targeting controls), using
In the next step, the median z-scores of individual siRNAs were
used to calculate a ranked gene list, using the redundant siRNA
analysis algorithm; genes with less than two hit-siRNAs (OPI-hits)
were excluded (RSA) . This list comprises genes leading to a
decreased p65 activity, when knocked down (termed inhibiting
For TNF-counter screening in HEK293T cells, as well as for hit
validation in HEK293T and THP1 cells, the top 435 inhibiting
hits were selected. For each of these genes, the two siRNAs
showing the strongest effect in the screen were re-synthezised and
assembled on 384-well plates (validation plates).
To validate the results of the primary screen, the experiments
were repeated as described above. Data analysis using CellHTS2
was done as described above; siRNAs were selected as inhibiting
Tri-DAP-hits, if their median Z-score exceeded the median of
non-targeting controls by more than two standard deviations.
To exclude unspecific hits, all siRNAs selected for validation
were screened for their influence on TNF-induced p65 activation.
Data analysis was done analogous to the Tri-DAP validation
screen. Inhibiting TNF-hits were excluded from further analysis.
Data from the THP1-blue screen consists of two parameters
(QUANTI-Blue absorption at 632 nm for Tri-DAP response
[QB], and XTT absorption at 485 nm for cell viability [XTT])
and was processed similar as described above: QB-signal of each
well was normalized to cell viability (XTT), yielding nQB
(normalized QUANTI-Blue absorption; QB/XTT = nQB). After
quality control and outlier flagging, the three best experimental
replicates were NPI-normalized to non-targeting and
NOD1control-siRNAs using CellHTS2, and median Z-scores were used
for hit identification. All genes with two siRNAs showing a
decrease of .1.5-fold standard deviation compared to the
nontargeting control were regarded as validated inhibiting hits. A
subset of these with one siRNA showing a decrease of .3.0-time
standard deviation were categorized as strong inhibiting hits (8
Gene accession numbers
SSH1: Gene ID 54434; NOD1: Gene ID 10392; NOD2: Gene
ID 64127; ROCK1: Gene ID 6093; ROCK2: Gene ID 9475.
Figure S1 Flow-chart and graphical illustration of the
basic screening protocol used in HEK293T cells. (A)
Layout of the screening procedure. (B) Schematic representation of
the NOD1 signaling pathway. Related to Figure 1.
Figure S2 Analysis of the screen data. (A) Distribution of
the non-targeting controls (green), positive control (p65, orange)
and all tested siRNAs (blue). Left panel: Dot-plot. Right panel:
Ranked presentation of all hits. (B) Gene ontology (GO)
enrichment of the 435 preliminary hit genes retrieved by the
primary screen compared to the druggable-genome library
(background). Number of hit genes associated with enriched terms
according to the GO of biological processes. Enrichment factors
and negative log(P-values) are indicated in brackets. (C) Known
regulators of NF-kB retrieved by the screens. Ingenuity Pathway
analysis of 56 (out of 435) preliminary inhibiting screen hits
already known to be involved in NF-kB regulation. Among them,
28 could be validated in the HEK validation screen (dark red:
strong hit; light red: weaker hit; white: not validated). A sub group
of 14 were not influencing TNF-a-induced NF-kB activation (blue
borders). (D) Distribution of the siRNAs in the THP1-blue
validation screen. Each dot represents the median Z-score of the
experimental replicates after NPI-normalisation (Normalized
Percent Inhibition) to the plate internal NOD1-controls. A
positive value represents an inhibitory effect of the respective
siRNA on NF-kB activation. THP1 hit-siRNAs are defined as
exceeding 1.5 S.D. of CTRL from the median of the CTRL
siRNAs. Z-scores were normalized to control siRNAs set to 0.
CTRL: non-targeting controls. (E) STRING database analysis of
the 28 hits from the THP1 screen. Genes, where the NPI values of
both siRNAs exceeded 1.5 S.D. of CTRL from the median of
CTRL were regarded as validated. Stronger associations are
represented by thicker blue lines. Genes were clustered according
to the Marcov cluster (MCL) algorithm. Related to Figure 1.
Figure S3 Validation of XIAP as essential component of
the NOD1 signaling pathway. (A) HEK293T cells were
transfected with CTRL or XIAP siRNA and incubated for 48 h.
Subsequently, cells were transfected with the NF-kB luciferase
RPS- containing NOD1 (left panel) or NOD2 (right panel)
expression plasmids and stimulated with TriDAP (0.5 mM) or
MDP (50 nM), respectively. After 16 h, cells were lysed, luciferase
activation was determined and normalized with the
b-galactosidase values (nRLU). Values are mean + S.D. (n = 3). (B)
Confirmation of the XIAP knock-down on protein level.
HEK293T cells were transfected with CTRL or XIAP siRNA.
After 72 h, cells were lysed and XIAP protein levels were
determined by Western blot using a XIAP-specific antibody.
Detection of GAPDH served as loading control. (C) Differentiated
THP1-blue cells were transfected with CTRL, NOD1, NOD2, or
XIAP siRNA. Cells were incubated for 72 h and subsequently
stimulated with TriDAP or MDP (10 mg/ml each), 16 h later,
NFkB activation was determined by measurement of SEAP activity in
the supernatants. (D) In parallel, IL-8 secretion was determined by
ELISA. Values are NF-kB activation [OD620] or IL-8 secretion
[pg/ml], respectively, mean +SD (CTRL, XIAP: n = 3; NOD1,
NOD2: n = 2). Data is representative for at least three independent
experiments. Related to Figure 1.
Figure S4 SSH1 knock-down effect in THP1 cells. PMA
differentiated THP1 cells were treated for 72 h with a
nontargeting (siCTRL) or two SSH1 specific siRNA duplex. (A) IL-8
secretion was measured by ELISA after stimulation for 16 h with
TriDAP (10 mg/ml), MDP (10 mg/ml), TNF (0.1 mg/ml), and or
LPS (0.05 mg/ml) stimulation. (B) In parallel, SSH1 mRNA levels
were determined by qPCR from the same cells. Values are Mean +
S.D. (n = 3) representative of at least three independent
experiments. * p,0.05. n.s.: not significant. (C) Cytokine release of 20
relevant inflammatory cytokines of the cells in (A) after TriDAP
stimulation was analyzed by multiplex bead array. Changes in the
cytokine response following SSH1 depletion by siRNA is shown
for the proteins with the highest induction. Related to Figure 2.
Figure S5 SSH1 knock-down phenotype in HeLa and
primary human dermal fibroblasts. (A) qPCR analysis of
SSH1 levels in the experiment shown in Figure 3C. (B)
Immunoblotting for cofilin S3-phosphorlyation in HEK293T cells
after SSH1 knock-down. (C) Primary human dermal fibroblasts of
early passages were transfected with the indicated siRNA for 72 h
and subsequently treated with TNF, TriDAP, or infected with
S.flexneri and IL-8 release was measured 6 h later by ELISA
(upper panel). Percent IL-8 level compared to cells treated with the
siCTRL SSH1 knock-down is shown by qPCR from the same cells
(lower panel). Mean + S.D. (n = 3) representative of experiments
from cells of two different donors is shown. Related to Figure 3.
Figure S6 SSH1 co-localizes with Nod1. (A) Indirect
immunofluorescence images of HeLa cells expressing
GFPNOD1 and myc-SSH1. NOD1 signal is shown in green in the
merge image together with the SSH1 signal in red and DNA
staining in blue. (B) HCS/HTS based PLA spot quantification.
Images show cell segmentation with HCS cell mask stain (blue
channel). Cell classification in EGFP-SSH1 positive/negative
(green channel, indicated in blue channel) and automated
PLAspot detection (red channel). Scale bar 25 mm. (C) Quantification
of PLA signals in cells expressing Flag-NOD1 and the indicated
SSH1 construct. The graph shows the mean 6 SEM of two
independent experiments with more than 850 cells analyzed per
sample and experiment. (D) Quantification of IL-8 in the
supernatants of the cells analyzed in Fig. 4 panel F. Mean +
S.D. of triplicates is shown. Related to Figure 4.
Figure S7 SSH1 does not affect the sub-cellular
localization of NOD1 and NOD2. (A) Indirect immunofluorescence
micrographs of HeLa cells transiently transfected with Flag-tagged
NOD1 or NOD2. The cells were treated with non-targeting
siRNA or a SSH1 specific siRNA where indicated. (B)
Knockdown efficiency is shown by immunoblotting. Probing for tubulin
served as loading control.
Figure S8 SSH1 phosphatase activity is involved in
Nod1 signaling. (A) TNF-induced NF-kB activity (filled bars)
measured by luciferase reporter assays in HEK293T cells for 6 h
with the ROCK inhibitors Y-27632 and Glycyl-H1152. The effect
on basal NF-kB activity is shown by open bars. Values are mean +
S.D. (n = 3). * p,0.05, ** p,0.005 (two-sided students t-test,
compared to mock treated cells 0). (B) Over expression of SSH1
in HEK293T cells. NF-kB activity measured by luciferase reporter
assays. Values are mean + S.D. (n = 3). (C) Effect of the expression
of a phosphatase-dead mutant (C393S) of SSH1 on
NOD1mediated NF-kB activity measured by luciferase reporter assays in
HEK293T cells. TriDAP induced cells are shown as filled bars,
untreated cells only expressing Nod1 are shown as open bars.
Values are mean + S.D. (n = 3). Related to Figure 5.
Table S1 Screen data. (A) Gene-based list of the inhibiting hits
from the HEK293T siRNA-screen. Genes are ranked according to
the logP-values derived by applying the RSA-algorithm to the
median Z-scores of the four individual siRNAs per gene. (B)
Genebased list of the hits validated in HEK293T cells. Genes are
ranked according to the median Z-score of the strongest siRNA.
Genes specifically involved in TriDAP-induced NF-kB activation
are shown in bold. (C) siRNA-based list of the data obtained from
the three secondary screens validation and counter screen using
HEK293T cells, as well as the THP1 validation screen.
We thank Martin Kronke for continuous support, Rike Zietlow (MPI-IB,
Berlin) for critical reading of the manuscript and Anna Damm (University
of Cologne) for help with experiments.
Conceived and designed the experiments: TAK AH NM TFM. Performed
the experiments: HB KL MM CB SAE SB. Analyzed the data: TAK AH
PRB NM TFM. Contributed reagents/materials/analysis tools: MA HK
PJS BZ. Wrote the paper: TAK AH HB KL.
1. Akira S , Uematsu S , Takeuchi O ( 2006 ) Pathogen recognition and innate immunity . Cell 124 : 783 - 801 .
2. Moreira LO , Zamboni DS ( 2012 ) NOD1 and NOD2 Signaling in Infection and Inflammation . Front Immunol 3 : 328 .
3. Ting JP , Duncan JA , Lei Y ( 2010 ) How the noninflammasome NLRs function in the innate immune system . Science 327 : 286 - 290 .
4. Chamaillard M , Hashimoto M , Horie Y , Masumoto J , Qiu S , et al. ( 2003 ) An essential role for NOD1 in host recognition of bacterial peptidoglycan containing diaminopimelic acid . Nat Immunol 4 : 702 - 707 .
5. Girardin SE , Boneca IG , Carneiro LA , Antignac A , Jehanno M , et al. ( 2003 ) Nod1 detects a unique muropeptide from gram-negative bacterial peptidoglycan . Science 300 : 1584 - 1587 .
6. Girardin SE , Tournebize R , Mavris M , Page AL , Li X , et al. ( 2001 ) CARD4/ Nod1 mediates NF-kappaB and JNK activation by invasive Shigella flexneri . EMBO Rep 2 : 736 - 742 .
7. Krieg A , Correa RG , Garrison JB , Le Negrate G , Welsh K , et al. ( 2009 ) XIAP mediates NOD signaling via interaction with RIP2 . Proc Natl Acad Sci U S A 106 : 14524 - 14529 .
8. Bertrand MJ , Doiron K , Labbe K , Korneluk RG , Barker PA , et al. ( 2009 ) Cellular inhibitors of apoptosis cIAP1 and cIAP2 are required for innate immunity signaling by the pattern recognition receptors NOD1 and NOD2 . Immunity 30 : 789 - 801 .
9. Kufer TA , Kremmer E , Adam AC , Philpott DJ , Sansonetti PJ ( 2008 ) The pattern-recognition molecule Nod1 is localized at the plasma membrane at sites of bacterial interaction . Cell Microbiol 10 : 477 - 486 .
10. Keestra AM , Winter MG , Auburger JJ , Frassle SP , Xavier MN , et al. ( 2013 ) Manipulation of small Rho GTPases is a pathogen-induced process detected by NOD1 . Nature 496 : 233 - 237 .
11. Fukazawa A , Alonso C , Kurachi K , Gupta S , Lesser CF , et al. ( 2008 ) GEF-H1 mediated control of NOD1 dependent NF-kappaB activation by Shigella effectors . PLoS Pathog 4 : e1000228 .
12. Eitel J , Krull M , Hocke AC , N'Guessan PD , Zahlten J , et al. ( 2008 ) Beta-PIX and Rac1 GTPase mediate trafficking and negative regulation of NOD2 . J Immunol 181 : 2664 - 2671 .
13. Legrand-Poels S , Kustermans G , Bex F , Kremmer E , Kufer TA , et al. ( 2007 ) Modulation of Nod2-dependent NF-kappaB signaling by the actin cytoskeleton . J Cell Sci 120 : 1299 - 1310 .
14. Lipinski S , Grabe N , Jacobs G , Billmann-Born S , Till A , et al. ( 2012 ) RNAi screening identifies mediators of NOD2 signaling: implications for spatial specificity of MDP recognition . Proc Natl Acad Sci U S A 109 : 21426 - 21431 .
15. Mizuno K ( 2013 ) Signaling mechanisms and functional roles of cofilin phosphorylation and dephosphorylation . Cell Signal 25 : 457 - 469 .
16. Zurek B , Bielig H , Kufer TA ( 2011 ) Cell-based reporter assay to analyze activation of Nod1 and Nod2 . Methods Mol Biol 748 : 107 - 119 .
17. Konig R , Chiang CY , Tu BP , Yan SF , DeJesus PD , et al. ( 2007 ) A probabilitybased approach for the analysis of large-scale RNAi screens . Nat Methods 4 : 847 - 849 .
18. Damgaard RB , Nachbur U , Yabal M , Wong WW , Fiil BK , et al. ( 2012 ) The ubiquitin ligase XIAP recruits LUBAC for NOD2 signaling in inflammation and innate immunity . Mol Cell 46 : 746 - 758 .
19. Kim ML , Jeong HG , Kasper CA , Arrieumerlou C ( 2010 ) IKKalpha contributes to canonical NF-kappaB activation downstream of Nod1-mediated peptidoglycan recognition . PLoS One 5 : e15371 .
20. Warner N , Burberry A , Franchi L , Kim YG , McDonald C , et al. ( 2013 ) A genome-wide siRNA screen reveals positive and negative regulators of the NOD2 and NF-kappaB signaling pathways . Sci Signal 6 : rs3 .
21. Huang TY , DerMardirossian C , Bokoch GM ( 2006 ) Cofilin phosphatases and regulation of actin dynamics . Curr Opin Cell Biol 18 : 26 - 31 .
22. Kurita S , Watanabe Y , Gunji E , Ohashi K , Mizuno K ( 2008 ) Molecular dissection of the mechanisms of substrate recognition and F-actin-mediated activation of cofilin- phosphatase Slingshot-1. J Biol Chem 283 : 32542 - 32552 .
23. Yamamoto M , Nagata-Ohashi K , Ohta Y , Ohashi K , Mizuno K ( 2006 ) Identification of multiple actin-binding sites in cofilin-phosphatase Slingshot-1L . FEBS Lett 580 : 1789 - 1794 .
24. Dai S , Sarmiere PD , Wiggan O , Bamburg JR , Zhou D ( 2004 ) Efficient Salmonella entry requires activity cycles of host ADF and cofilin . Cell Microbiol 6 : 459 - 471 .
25. Olayioye MA , Barisic S , Hausser A ( 2013 ) Multi-level control of actin dynamics by protein kinase D . Cell Signal25 : 1739 - 1747 .
26. Kustermans G , El Benna J , Piette J , Legrand-Poels S ( 2005 ) Perturbation of actin dynamics induces NF-kappaB activation in myelomonocytic cells through an NADPH oxidase-dependent pathway . Biochem J 387 : 531 - 540 .
27. Kustermans G , El Mjiyad N , Horion J , Jacobs N , Piette J , et al. ( 2008 ) Actin cytoskeleton differentially modulates NF-kappaB-mediated IL-8 expression in myelomonocytic cells . Biochem Pharmacol 76 : 1214 - 1228 .
28. Abbott DW , Yang Y , Hutti JE , Madhavarapu S , Kelliher MA , et al. ( 2007 ) Coordinated regulation of Toll-like receptor and NOD2 signaling by K63-linked polyubiquitin chains . Mol Cell Biol 27 : 6012 - 6025 .
29. Hasegawa M , Fujimoto Y , Lucas PC , Nakano H , Fukase K , et al. ( 2008 ) A critical role of RICK/RIP2 polyubiquitination in Nod-induced NF-kappaB activation . EMBO J 27 : 373 - 383 .
30. Bianchi K , Meier P ( 2009 ) A tangled web of ubiquitin chains: breaking news in TNF-R1 signaling . Mol Cell 36 : 736 - 742 .
31. Hayden MS , Ghosh S ( 2008 ) Shared principles in NF-kappaB signaling . Cell 132 : 344 - 362 .
32. Andree M , Seeger JM , Schull S , Coutelle O , Wagner-Stippich D , et al. ( 2014 ) BID-dependent release of mitochondrial SMAC dampens XIAP-mediated immunity against Shigella . EMBO J . in press
33. Yeretssian G , Correa RG , Doiron K , Fitzgerald P , Dillon CP , et al. ( 2011 ) Nonapoptotic role of BID in inflammation and innate immunity . Nature 474 : 96 - 99 .
34. Regueiro V , Moranta D , Frank CG , Larrarte E , Margareto J , et al. ( 2011 ) Klebsiella pneumoniae subverts the activation of inflammatory responses in a NOD1-dependent manner . Cell Microbiol 13 : 135 - 153 .
35. Arbibe L , Mira JP , Teusch N , Kline L , Guha M , et al. ( 2000 ) Toll-like receptor 2-mediated NF-kappa B activation requires a Rac1-dependent pathway . Nat Immunol 1 : 533 - 540 .
36. Porter K , Shimono M , Tian M , Day B ( 2012 ) Arabidopsis Actin-Depolymerizing Factor-4 links pathogen perception, defense activation and transcription to cytoskeletal dynamics . PLoS Pathog 8 : e1003006 .
37. Menning M , Kufer TA ( 2013 ) A role for the Ankyrin repeat containing protein Ankrd17 in Nod1- and Nod2-mediated inflammatory responses . FEBS Lett 587 : 2137 - 2142 .
38. Zigrino P , Drescher C , Mauch C ( 2001 ) Collagen-induced proMMP-2 activation by MT1-MMP in human dermal fibroblasts and the possible role of alpha2beta1 integrins . Eur J Cell Biol 80 : 68 - 77 .
39. Barisic S , Nagel AC , Franz-Wachtel M , Macek B , Preiss A , et al. ( 2011 ) Phosphorylation of Ser 402 impedes phosphatase activity of slingshot 1 . EMBO Rep 12 : 527 - 533 .
40. Ishizaki T , Naito M , Fujisawa K , Maekawa M , Watanabe N , et al. ( 1997 ) p160ROCK, a Rho-associated coiled-coil forming protein kinase, works downstream of Rho and induces focal adhesions . FEBS Lett 404 : 118 - 124 .
41. Kufer TA , Kremmer E , Banks DJ , Philpott DJ ( 2006 ) Role for erbin in bacterial activation of Nod2 . Infect Immun 74 : 3115 - 3124 .
42. Boutros M , Bras LP , Huber W ( 2006 ) Analysis of cell-based RNAi screens . Genome Biol 7 : R66 .