Sensing low intracellular potassium by NLRP3 results in a stable open structure that promotes inflammasome activation
Abstract
El inflamasoma NLRP3 se activa por una amplia gama de estímulos y conduce a diversas enfermedades inflamatorias. La disminución de la concentración de K+ intracelular es una señal mínima en la mayoría de los modelos de activación del NLRP3. Aquí, encontramos que el eflujo de K+ celular induce un cambio estructural estable en el NLRP3 inactivo, promoviendo una conformación abierta como paso previo a la activación. Este cambio conformacional se ve facilitado por el dominio específico FISNA de NLRP3 y una secuencia única de enlace flexible entre los dominios PYD y FISNA. Este enlazador también facilita el ensamblaje de NLRP3PYD en una estructura semilla para la oligomerización de ASC. La introducción de la secuencia enlazadora PYD-FISNA de NLRP3 en NLRP6 dio lugar a un receptor quimérico capaz de ser activado por activadores de NLRP3 específicos del eflujo de K+ y promovió una respuesta inflamatoria in vivo a los cristales de ácido úrico. Nuestros resultados establecen que la secuencia aminoterminal entre el dominio PYD y NACHT de NLRP3 es clave para la activación del inflamasoma.
The
NLRP3 inflammasome is activated by a wide range of stimuli and drives
diverse inflammatory diseases. The decrease of intracellular K+ concentration is a minimal upstream signal to most of the NLRP3 activation models. Here, we found that cellular K+
efflux induces a stable structural change in the inactive NLRP3,
promoting an open conformation as a step preceding activation. This
conformational change is facilitated by the specific NLRP3 FISNA domain
and a unique flexible linker sequence between the PYD and FISNA domains.
This linker also facilitates the ensemble of NLRP3PYD into a
seed structure for ASC oligomerization. The introduction of the NLRP3
PYD-linker-FISNA sequence into NLRP6 resulted in a chimeric receptor
able to be activated by K+ efflux–specific NLRP3 activators
and promoted an in vivo inflammatory response to uric acid crystals. Our
results establish that the amino-terminal sequence between PYD and
NACHT domain of NLRP3 is key for inflammasome activation.
INTRODUCTION
El NLRP3 forma un inflamasoma único porque puede activarse en respuesta a
muchos estímulos diferentes y exclusivos; la mayoría de ellos son
patrones moleculares estériles asociados a daños (DAMP) del huésped (1,
2). Esto implica a NLRP3 en la fisiopatología de diferentes enfermedades
inflamatorias crónicas que no son impulsadas por infecciones, así como
en enfermedades metabólicas y neurodegenerativas (2). Los diferentes
estímulos que activan el NLRP3 convergen en la interacción del NLRP3 con
varias moléculas accesorias para facilitar su activación. Estas
moléculas incluyen, entre otras, la proteína nunca en la mitosis
A-relacionada con la quinasa 7 (NEK7), la proteína que interactúa con la
tiorredoxina (TXNIP), la quinasa reguladora de la afinidad de los
microtúbulos 4 (MARK4), la proteína ligada al estrés DEAD box helicase 3
X-linked (DDX3X), la proteína mitocondrial de señalización antiviral
(MAVS) y el receptor de la proteína C quinasa 1 activada (RACK1), así
como la cardiolipina expuesta en las mitocondrias o el
fosfatidilinositol-4-fosfato (PtdIns4P) en la red trans-Golgi dispersa
(3-9). Sin embargo, la disminución del K+ intracelular es un paso
celular mínimo común para la mayoría de los estímulos de NLRP3 (10, 11),
siendo la interacción de NLRP3 con NEK7 o con PtdIns4P en la red
trans-Golgi dispersa dependiente del flujo de K+ intracelular (3, 4).
Sin embargo, aún se desconoce cómo una disminución de la concentración
de K+ intracelular, un paso común clave para la activación de NLRP3
(12), favorece el ensamblaje del inflamasoma NLRP3. Tras la activación,
NLRP3 se homo-oligomiza, formando una estructura multimérica que recluta
a la proteína adaptadora ASC (13), y luego las moléculas de ASC forman
grandes filamentos mediante las subsiguientes interacciones homotípicas
ASCPYD-ASCPYD (14). A continuación, los filamentos de ASC forman una
única estructura de gran tamaño (denominada "speck" de ASC), que recluta
y activa la caspasa-1 (15, 16). La caspasa-1 activa controla la
señalización del inflamasoma aguas abajo ejecutando el procesamiento de
citoquinas proinflamatorias, como la interleucina (IL)-1β, y la
gasdermina D, una proteína que induce la muerte celular piroptótica
(17). La piroptosis también permite la liberación de motas de ASC,
amplificando la respuesta inflamatoria y provocando, entre otros
efectos, el depósito de amiloide (18-20).
NLRP3
forms a unique inflammasome because it can be activated in response to
many different and exclusive stimuli; most of them are sterile
damage-associated molecular patterns (DAMPs) from the host (1, 2).
This implicates NLRP3 in the pathophysiology of different chronic
inflammatory diseases that are not driven by infections, as well as in
metabolic and neurodegenerative diseases (2).
The different NLRP3-activating stimuli converge in the interaction of
NLRP3 with several accessory molecules to facilitate its activation.
These molecules include, among others, protein never in mitosis
A–related kinase 7 (NEK7), thioredoxin-interacting protein (TXNIP),
microtubule-affinity regulating kinase 4 (MARK4), stress granule DEAD
box helicase 3 X-linked protein (DDX3X), mitochondrial
antiviral-signaling protein (MAVS), and receptor for activated protein C
kinase 1 (RACK1), as well as exposed cardiolipin in mitochondria or
phosphatidylinositol-4-phosphate (PtdIns4P) in dispersed trans-Golgi
network (3–9). However, the decrease of intracellular K+ is a common minimal cellular step for most of the NLRP3 stimuli (10, 11), being the interaction of NLRP3 with NEK7 or with PtdIns4P in dispersed trans-Golgi network dependent on intracellular K+ efflux (3, 4). However, how a decrease in the intracellular K+ concentration, a key common step for NLRP3 activation (12),
favors NLRP3 inflammasome assembly remains unknown. After activation,
NLRP3 homo-oligomerizes, forming a multimeric structure that recruits
the adaptor protein ASC (13), and then ASC molecules form large filaments by subsequent ASCPYD-ASCPYD homotypic interactions (14). Then, ASC filaments form a single large structure (called ASC “speck”), which recruits and activates caspase-1 (15, 16).
Active caspase-1 controls downstream inflammasome signaling executing
the processing of proinflammatory cytokines, such as interleukin
(IL)–1β, and gasdermin D, a protein inducing pyroptotic cell death (17).
Pyroptosis also allows the release of ASC specks, amplifying the
inflammatory response and provoking, among other effects, amyloid
deposition (18–20).
In
this study, by using a bioluminescence resonance energy transfer (BRET)
technique to monitor NLRP3 conformational changes, we have identified
that the specific N-terminal NLRP3 linker and FISNA (fish-specific NACHT
associated) domain are key to provoke a stable change on the inactive
NLRP3 structure after a decrease of intracellular K+
concentration, allowing activation of NLRP3. The sequence between the
PYD and FISNA domains should probably be a flexible linker that also
favors the correct placement of NLRP3PYD to engage ASC into the inflammasome. However, this linker sequence is dispensable for K+ efflux–dependent NLRP3 structural change and oligomerization, as NLRP3 with deletions of this region are activated in a K+
efflux–dependent manner. NLRP6 harboring the NLRP3 PYD-linker-FISNA
sequence is able to form a functional inflammasome in response to K+
efflux and drives an in vivo immune response in the NLRP3-specific uric
acid crystal model. These different unique structural features of NLRP3
involved in this response can explain why only NLRP3 and no other NLRs
can be activated in response to intracellular K+ decrease.
RESULTS
NLRP3 structure is modified by intracellular K+ efflux
Lipopolysaccharide (LPS)–primed mouse macrophages treated with increased concentrations of the K+ ionophore nigericin dose-dependently decreased intracellular K+ concentration and, in parallel, increased IL-1β release [calculated median inhibitory concentration (IC50) of 0.8 μM and 1.5 μM, respectively] (Fig. 1A). Similarly, other two K+ ionophores, valinomycin and BB15C1, were also able to induce IL-1β release with a parallel decrease of intracellular K+
concentration (fig. S1). Nigericin-induced IL-1β release was
dose-dependently inhibited by increasing concentrations of extracellular
KCl, with an IC50 of 16 mM (Fig. 1B). To rule out unspecific effects of the increase of extracellular K+ concentrations, we used extracellular Rb+, a slightly larger metal ion analog to K+ that mimics a lower K+
conductance in most ion channels and ionophores. We found that
extracellular RbCl, but not CsCl or LiCl (ions that do not permeate
through K+-selective pores), blocked IL-1β release induced by nigericin, without significantly impairing IL-1β release associated to the K+-insensitive Pyrin inflammasome activated by Clostridium difficile toxin B (TcdB) (Fig. 1C). These results and previous data (12, 21) confirm that a decrease of intracellular K+ is linked to the activation of the NLRP3 inflammasome. To gain insight on how the decrease in intracellular K+
could activate NLRP3, we used human embryonic kidney (HEK) 293T cells
as a validated cellular model to study NLRP3 oligomerization and
activation (4, 22–24). We found that nigericin treatment induced NLRP3 oligomerization, being this oligomerization blocked by extracellular KCl (Fig. 1D).
NLRP3 oligomers induced by nigericin in HEK293T cells colocalize with
ASC specks (when ASC was coexpressed together NLRP3), similarly to when
NLRP3 is activated in macrophages (Fig. 1E),
suggesting that the NLRP3 oligomers found in HEK293T cells after
nigericin treatment present a functional structure. We next used an
NLRP3 BRET sensor expressed in HEK293T cells (22, 25)
and found that nigericin induced a decrease of the BRET signal for
NLRP3 that was stable for up to 1.5 hours after stimulation (Fig. 2A). Increasing extracellular KCl concentration impaired nigericin-induced NLRP3 BRET signal change (Fig. 2A), indicating that K+
efflux could stably change the structure of NLRP3. The structural
change induced by nigericin in NLRP3 was dose-dependently inhibited by
increasing concentrations of extracellular KCl with an IC50 of 43 mM (Fig. 2B). Extracellular Rb+-based buffer, but not a Li+-based buffer, also blocked the stable NLRP3 BRET signal decrease induced by nigericin (Fig. 2C). The use of valinomycin and BB15C1 K+
ionophores, as well as the activation of P2X7 receptor with adenosine
triphosphate (ATP) (in HEK293T cells expressing P2X7), resulted in a
stable decrease of NLRP3 BRET signal that was impaired when the
extracellular solution contained elevated concentration of KCl (Fig. 2D).
The decrease of the NLRP3 BRET signal has been recently reported to be a
transient intermediate NLRP3 structure with an open conformation that
promotes activation (22), indicating that intracellular K+
efflux could open NLRP3 structure favoring its oligomerization. In the
absence of a full-length NLRP3 structure in inactive and active
conformation, it is difficult to precisely interpret the change in BRET
signal. Our previous studies demonstrated that upon NLRP3 triggering,
the NLRP3 BRET signal is reduced and that NLRP3 with CAPS-related
mutants have lower resting BRET signals (9, 24–27),
indicating a distancing of luciferase and yellow fluorescent protein
(YFP) epitopes by an opening of the NLRP3 structure. Therefore, we
modeled a semi-open and an open NLRP3 conformation based on the recent
NLRP3 structure (6NPY) and the structure of NLRC4 inside an oligomer
(3JBL) (28, 29),
and we found that structurally a hinge that allowed this conformational
change was on the NACHT domain (residues 417–441 inside the HD1 motif;
fig. S2A). This hinge was able to move in block a compact helix bundle
of the NACHT domain and was conserved in the NACHT of other NLRs, as
NLRP6 (fig. S2A). By analyzing NLRP3 structures, we found that in the
semi-activated not fully open structure (6NPY), the nucleotide binding
pocket was hidden and would prevent ATP entry (fig. S2B). On the
contrary, the open NLRP3 structure presents an accessible nucleotide
binding pocket that could allow exchange of nucleotides and the entry of
ATP (fig. S2B). Because ATP hydrolysis is important for NLRP3
activation (30),
our data support the idea that the conformational opening of NLRP3
structure would favor ATP entry and further NLRP3 activation.
Fig. 1. K+ efflux induces cells with multiple NLRP3 puncta.
(A) IL-1β release and intracellular K+ concentration from LPS-primed bone marrow–derived macrophages (BMDM) activated for 30 min with different doses of nigericin. n = 4 to 6 independent experiments for each concentration of nigericin. (B)
IL-1β release from LPS-primed BMDMs activated for 30 min with nigericin
(10 μM) in the presence of increasing concentrations of KCl. (C) IL-1β release from LPS-primed BMDMs activated for 30 min with nigericin (10 μM, top) or with C. difficile toxin B (TcdB; 1 μg/ml, bottom) in the absence/presence of 40 mM KCl, RbCl, CsCl, or LiCl. n = 4 independent experiments; Kruskal-Wallis test. (D)
Representative fluorescent micrographs of human embryonic kidney (HEK)
293T cells expressing YFP-NLRP3 (green) and stained for nuclei
[4′,6-diamidino-2-phenylindole (DAPI); blue] treated for 30 min with
nigericin (10 μM) in the absence/presence of 140 mM KCl. Scale bar, 10
μm. (E) Representative fluorescent micrographs of HEK293T cells (left) and Nlrp3−/−
immortalized macrophages (iMos; right) expressing YFP-NLRP3 (green) and
ASC (stained in red) and stained for nuclei (DAPI, blue) and treated
for 30 and 60 min with nigericin (10 μM), respectively. Scale bar, 20 μm
(left) and 10 μm (right). Arrows denote ASC specks; HEK293T cells were
stably expressing YFP-NLRP3 and transfected for ASC.
Fig. 2. K+ efflux modifies the NLRP3 structure.
(A)
Average BRET signal for YFP-NLRP3-Luc expressed in HEK293T cells before
and after stimulation with nigericin (10 μM, denoted by a gray box) in
the absence (left, middle) or presence (right) of 140 mM KCl; note that
kinetics of BRET have been plotted with the same scale, but relatively
at the same level to ease comparison. n = 51, 4, and 6 for nigericin 15 min, nigericin 85 min, and nigericin + KCl, respectively; Mann-Whitney test. (B)
NLRP3 BRET signal variation after treatment for 15 min with nigericin
(10 μM) versus basal conditions without nigericin in the presence of
increasing concentrations of KCl. n = 3 to 5 independent cell culture wells for each KCl concentration. (C)
Average BRET signal for YFP-NLRP3-Luc expressed in HEK293T cells before
and after stimulation with nigericin (10 μM) in the absence/presence of
40 mM RbCl or 40 mM LiCl; note that kinetics of BRET have been plotted
with the same scale, but relatively at the same initial level to ease
comparison. n = 12 independent cultures from four independent experiments; Mann-Whitney test. (D)
Average BRET signal for YFP-NLRP3-Luc expressed in HEK293T cells (top
and middle) or HEK293T-P2X7 receptor (bottom), before and after
stimulation with valinomycin (50 μM), BB15C5 (50 μM), or ATP (3 mM), in
the absence/presence of 140 mM KCl; note that kinetics of BRET have been
plotted with the same scale, but relatively at the same initial level
to ease comparison. n = 6 to 18 independent cell cultures; Mann-Whitney test.
NLRP3PYD is dispensable for NLRP3 oligomerization but is necessary to form active NLRP3 oligomers
We recently demonstrated that the LRR domain of NLRP3 is not a repressor domain and is dispensable for NLRP3 activation (31);
this led us to study whether the N-terminal domain could have a role in
the activation of NLRP3. NLRP3 lacking the PYD N-terminal domain
(ΔPYD-NLRP3, Δ1–91) was able to oligomerize after nigericin stimulation (Fig. 3A) and to reduce the BRET signal similar to the one observed in NLRP3 wild type after nigericin treatment (Figs. 2A and 3B). ΔPYD-NLRP3 BRET signal was intramolecular (fig. S3A), similarly to the full-length NLRP3 (22, 25).
Oligomerization and the decrease of the ΔPYD-NLRP3 BRET signal induced
by nigericin were blocked using an extracellular buffer with elevated
KCl concentration (Fig. 3, A and B), demonstrating that the PYD domain is not necessary for NLRP3 oligomerization in response to intracellular K+
efflux. However, ΔPYD-NLRP3 expressed in NLRP3-deficient macrophages
(fig. S3B) was unable to induce the release of IL-1β after nigericin
activation (Fig. 3C), as it could not bind to ASC (Fig. 3D).
Therefore, the PYD domain is not necessary for NLRP3 oligomerization,
suggesting that it is not the trigger-sensing domain, but is crucial to
form fully functional NLRP3 oligomers by allowing the recruitment of
ASC.
Fig. 3. The PYD domain is dispensable for NLRP3 activation in response to K+ efflux.
(A)
Representative fluorescent micrographs of HEK293T cells expressing
YFP-ΔPYD-NLRP3 (green) and stained for nuclei (DAPI, blue) treated for
30 min with nigericin (10 μM) in the absence/presence of 140 mM KCl.
Scale bar, 10 μm. (B) Average BRET signal for YFP-ΔPYD-NLRP3-Luc
expressed in HEK293T cells before and after stimulation with nigericin
(10 μM) in the absence/presence of 140 mM KCl. n = 5 to 17
independent cell cultures, from three to five independent experiments;
note that kinetics of BRET have been plotted with the same scale, but
relatively at the same initial level to ease comparison; Mann-Whitney
test. (C) IL-1β release from Nlrp3−/−
immortalized macrophages (iMos) treated for 16 hours with doxycycline (1
μg/ml) and LPS (100 ng/ml) to induce the expression of YFP-ΔPYD-NLRP3
and then activated for 60 min with nigericin (10 μM). n = 3 independent experiments. (D)
Representative fluorescent micrographs of HEK293T cells expressing
YFP-ΔPYD-NLRP3 or full-length NLRP3 as indicated (green) and ASC (red),
stained for nuclei (DAPI, blue) and treated for 30 min with nigericin
(10 μM). Scale bar, 20 μm.
The NLRP3 sequence between the PYD and NACHT is important for inflammasome activation in response to K+ efflux
Because the PYD domain was not critical for NLRP3 oligomerization in response to K+
efflux but was important to engage ASC and form functional
inflammasomes, we decided to model the NLRP3 oligomeric structure
including the N-terminal PYD and the sequence up to the NACHT domain. In
NLRP3, this sequence is encoded by a specific exon present only in
mammalian NLRP3 sequences among all NLRPs (fig. S4, A to C) and then is
followed by the FISNA domain [PFAM (protein family database): PF14484], a
domain frequently found in proteins associated with the NACHT domain
and only present in two of the human NLRs: NLRP3 and NLRP12 (fig. S4C).
NLRP3 modeling was done using the structure of NLRP3 in complex with
NEK7 (6NPY) (28) and the structure of the NLRP3PYD domain (3QF2) to complete the structure that was optimized to avoid clashes (Fig. 4, A and B). This model showed that the NLRP3FISNA domains were actually interacting between NLRP3 monomers inside the oligomer (Fig. 4, A and B).
The sequence encoded by the NLRP3 exon 3 (residues 92–132) appeared as a
linker formed by an α helix and a flexible sequence that is able to
position the PYD domains of the different NLRP3 subunits in a compatible
conformation with the assembly of ASCPYD domains forming an helical fiber (Fig. 4, A and B). At the end of the α helix of the linker, there is a polybasic sequence (KMKK132)
associated with NLRP3 activation, as it mediates the binding of NLRP3
to negatively charged phospholipids, as PtdIns4P in the dispersed
trans-Golgi network, a phenomenon occurring by a stimulus-specific
mechanism (4).
Our model predicts that these positively charged residues, together
with the positive residues at the beginning of the FISNA domain
(RKKYRKYVRSR145) (fig. S5), allow the linker α helix to
position the PYD domain nearer or farther the FISNA domain in the
oligomeric structure, placing it in an oligomeric helix and creating a
seed for ASCPYD nucleation (Fig. 4A).
On the basis of this model, we would expect that the removal of the
NLRP3 linker sequence would decrease the engagement of ASC; meanwhile,
the lack of the FISNA domain would markedly affect NLRP3
oligomerization.
Fig. 4. Model of ASCPYD-NLRP3-NEK7 oligomer.
(A) The structure on the left corresponds to a ribbon plot showing an oligomer complex of 11 NLRP3 monomers, 11 NEK7, and 11 ASCPYD
chains. The representation in the right shows the superposition of 11
NLRP3 monomers of the oligomer complex of the modeled inflammasome
structure. PYD domains are placed at different distances of the NACHT
domain depending on the conformation of the linker (shown in dark red in
the right representation). (B) Oligomer complex of 11 NLRP3 monomers, 11 NEK7, and 11 ASCPYD
chains. Linker fragment between PYD domains and FISNA domains of each
NLRP3 is shown in dark red. FISNA domains of NLRP3 are shown in
gray-blue, highlighting the continuous connectivity through FISNA-FISNA
and FISNA-NACHT interfaces. This model presents a compatible structure
to promote ASCPYD filament formation.
In Nlrp3−/−
macrophages, expression of NLRP3 serial truncations of the sequence
between the PYD and NACHT domain, maintaining the PYD domain (fig. S4A),
confirmed a progressive lack of activity as denoted by a decrease of
IL-1β release induced by nigericin (Fig. 5A).
Although the release of IL-1β induced by nigericin was reduced when
macrophages expressed NLRP3 Δ92–120 and Δ92–132 (lacking partially or
totally the linker sequence), it was blocked when nigericin was added in
an extracellular buffer with an elevated KCl concentration (Fig. 5A).
Expression of NLRP3 truncations getting into the FISNA domain (Δ92–148)
or deleting the middle or final sequence of the FISNA domain (Δ149–180
or Δ181–217), or complete sequence between PYD and NACHT (Δ92–217)
resulted in a receptor that was unable to induce IL-1β release after
nigericin stimulation (Fig. 5A and fig. S6A). The expression of NLRP3 Δ92–120 in Nlrp3−/− macrophages resulted in nigericin-induced activation of caspase-1 and processing of IL-1β and GSDMD (Fig. 5B). The specific NLRP3 inhibitor MCC950 was able to block caspase-1 activation and IL-1β release by NLRP3 Δ92–120 activation (Fig. 5B),
confirming that this NLRP3 truncation presented a canonical NLRP3
activation, because the target sequence for MCC950 relays on the NLRP3NACHT domain (22, 32).
In contrast, these macrophages expressing different NLRP3 truncations
released similar concentrations of IL-1β when the Pyrin inflammasome was
activated (fig. S6B) and similar concentrations of IL-6 after LPS
priming (fig. S6C). Our NLRP3 structural modeling predicts that the
linker sequence is required to position NLRP3PYD in an optimal helical filament that will seed and nucleate ASCPYD but not to stabilize NLRP3 oligomers (Fig. 4, A and B).
We confirmed that macrophages expressing NLRP3 Δ92–120 and Δ92–132
(lacking partially or totally the linker sequence) were able to increase
the percentage of cells with ASC specks when treated with nigericin,
but this increase was highly reduced when NLRP3 Δ92–132 was expressed
and completely absent when NLRP3 Δ92–148 was expressed (Fig. 5, C and D).
We then studied oligomerization of NLRP3 Δ92–120, Δ92–132, and Δ92–148
truncations when expressed in HEK293T cells (lacking ASC expression).
Both NLRP3 Δ92–120 and Δ92–132 truncations, but not the Δ92–148
truncation, were able to oligomerize after nigericin stimulation (Fig. 5D). As our model predicted (Fig. 4A), in macrophages expressing the NLRP3 Δ92–132 truncation, several NLRP3 oligomers were not colocalizing with the ASC specks (Fig. 5D).
These data support the idea that the FISNA domain would be important
for NLRP3 oligomerization in response to nigericin. However, the linker
sequence could be important to allow an optimal engagement of ASC to the
NLRP3 oligomers. Truncation of the linker sequence (92–132) resulted in
a less flexible linker, which did not affect NLRP3 oligomerization upon
K+ efflux. We next found out that the expression of NLRP3
BRET sensor carrying different deletions in the region between PYD and
NACHT domain in HEK293T cells resulted in an intramolecular BRET signal
(fig. S6, D and E). Nigericin induced a K+ efflux–dependent stable reduction of NLRP3 BRET signal for Δ92–120 and Δ92–132 deletions but not for Δ92–148 (Fig. 5E).
The BRET signal from different deletions of the FISNA sequence
(Δ149–180 and Δ181–217), which failed to induce IL-1β release in
response to nigericin (fig. S6A), was also not affected by nigericin (Fig. 5F). Moreover, the reduction of NLRP3 Δ92–132 BRET signal induced by nigericin was reverted in the presence of MCC950 (Fig. 5E), similarly to the effect of this compound on the full-length NLRP3 BRET signal (22).
This further supports the idea that the stable reduction of NLRP3 BRET
signal lacking the linker sequence reflect an open active NLRP3
conformation. The fact that human NLRP3 Δ92–132 was still able to
respond to K+ efflux induced by nigericin could suggest that the identified polybasic sequence in the mouse NLRP3 sequence (KKKK130) important for activation (4) might be functioning differentially in the human NLRP3. Specific mutations of the KMKK132 polybasic sequence of human NLRP3 to AMAA132 released similar IL-1β after nigericin stimulation when compared to NLRP3 wild-type or polybasic RMRR132 rescue mutant expressed in macrophages lacking endogenous wild-type NLRP3 (fig. S6F). High extracellular K+
blocked this activation, and the release of IL-1β dependent on Pyrin
inflammasome activation was unaffected by the presence of these
mutations (fig. S6F). Furthermore, NLRP3 AMAA132 mutant was also able to oligomerize after nigericin treatment when expressed in HEK293 cells in a K+ efflux–dependent manner (fig. S6G). Specific AMAA132
mutation in NLRP3 also resulted in a reduced BRET signal in response to
nigericin with a similar profile than wild-type and RMRR132
mutant NLRP3 (fig. S6H). These data suggest that binding to PtdIns4P in
the dispersed trans-Golgi network in the human NLRP3 by the KMKK132 sequence is not as critical as for the mouse NLRP3 with a KKKK130 sequence (4). This could reflect that the presence of a second polybasic region in the human NLRP3 (RKKYRKYVRSR145) at the beginning of the FISNA domain could be important for human NLRP3 interaction with dispersed trans-Golgi network after K+ efflux.
Fig. 5. The sequence between PYD and NATCH domains is important for NLRP3 activation in response to K+ efflux.
(A) IL-1β release from Nlrp3−/−
immortalized macrophages (iMos) treated for 16 hours with doxycycline
(1 μg/ml) and LPS (100 ng/ml) to induce the expression of different
truncations of NLRP3 as indicated and then activated for 60 min with
nigericin (10 μM) in the absence/presence of 40 mM KCl. n = 2 to 5 independent experiments; Mann-Whitney test. (B) Western blot for IL-1β, caspase-1, GSDMD, NLRP3, and β-actin from cell lysates and supernatants of Nlrp3−/− iMos expressing YFP-Δ92–120-NLRP3 treated as in (A) in the absence/presence of MCC950 (10 μM); representative of n = 2 independent experiments. (C) Percentage of ASC specking Nlrp3−/− iMos expressing different truncations of NLRP3 treated as in (A). n = 3 to 27 independent experiments; Mann-Whitney test. (D) Representative fluorescent micrographs of Nlrp3−/−
iMos (top) and HEK293T cells (bottom) expressing different truncations
of YFP-NLRP3 as indicated (green) and iMos stained for ASC [red,
arrowheads denote ASC specks quantified in (C)] and nuclei (DAPI, blue),
treated for 60 and 30 min with nigericin (10 μM), respectively. Scale
bar, 10 μm. Quantification of HEK293T cells with NLRP3 punctuate
staining is shown on the right of micrographs. n = 3 to 10 independent experiments. (E and F)
Average BRET signal for YFP-Δ92–120-NLRP3-Luc, YFP-Δ92–132-NLRP3-Luc,
YFP-Δ92–148-NLRP3-Luc (E), or YFP-Δ149-180-NLRP3-Luc,
YFP-Δ181–217-NLRP3-Luc (F) expressed in HEK293T cells before and after
stimulation with nigericin (10 μM, indicated by an arrow) in the
absence/presence of 140 mM KCl or MCC950 (10 μM); note that kinetics of
BRET have been plotted with the same scale, but relatively at the same
initial level to ease comparison. n = 2 to 19 independent cell cultures; Mann-Whitney test.
NLRP3 PYD-linker-FISNA sequence renders NLRP6 sensitive to K+ efflux
To
corroborate the structural model data and that the NLRP3 linker-FISNA
sequence is important to activate NLRP3 inflammasome in response to K+ efflux, we constructed chimeric receptors with NLRP6, a receptor that is not activated by K+ efflux (33, 34).
NLRP6 presents a shorter sequence between the PYD and NACHT domains
because it lacks the flexible linker sequence coded by the unique NLRP3
exon 3 (fig. S7, A and B). Furthermore, NLRP6 does not have a conserved
annotated FISNA domain but presents some conserved features with the
NLRP3 FISNA domain, particularly an initial polybasic sequence and some
structural motifs, including three putative residues interacting with
the nucleotide, with an overall similarity of 46.1% and a 49.3% homology
with the FISNA family signature (fig. S7, B and C). We first
demonstrated that NLRP6 was not endogenously expressed in the Nlrp3−/−
macrophages used in this study (fig. S8A) and the expression of NLRP6
led to the release of IL-1β induced by lipoteichoic acid (fig. S8B),
indicating that NLRP6, when expressed in these macrophages, was able to
form functional inflammasomes as has been already reported (33). We then expressed in Nlrp3−/− macrophages NLRP6 containing different lengths of the NLRP3 N-terminal sequence, starting from the NLRP3PYD domain (1–91), the NLRP3PYD
and flexible linker sequence (1–132), and the NLRP3 PYD-linker-FISNA
(1–217). Nigericin was not able to induce the release of IL-1β when
macrophages expressed the full-length NLRP6, but the NLRP3/6 chimeric
receptors released increased concentrations of IL-1β in response to
nigericin when the longer NLRP3 sequence was introduced into the NLRP6 (Fig. 6A). However, the K+
efflux dependence for IL-1β release after nigericin stimulation was
observed for the full-length NLRP3 and the chimera
NLRP3(1–217)-NLRP6(196–892) (Fig. 6A).
These data support the notion that the NLRP3 N-terminal sequence
between PYD and NACHT facilitates the formation of functional
inflammasomes that promote the recruitment of ASC in response to K+
efflux. Because of the similarity of NLRP3 FISNA with the NLRP6
sequence between PYD and NACHT (fig. S7B), this last one could be taking
a similar conformation and some functionality in sensing K+ efflux, because NLRP6 also present a polybasic sequence KKKYREHVLQL129 (fig. S7, A and B) that was present in the chimeras NLRP3(1–91)-NLRP6(104–892) and NLRP3(1–132)-NLRP6(104–892) (Fig. 6A).
Expression of the different chimeric receptors was not able to affect
IL-1β release after Pyrin inflammasome activation or IL-6 release after
LPS priming (fig. S8, C and D). Furthermore, the chimera
NLRP3(1–217)-NLRP6(196–892) was able to activate caspase-1 and process
IL-1β in response to nigericin in a caspase-1–dependent manner (Fig. 6, B and C), but the release of IL-1β induced by nigericin was insensitive to MCC950 (Fig. 6C), supporting the idea that this compound specifically targets the NLRP3NACHT, but not the NLRP6NACHT, domain (22, 32). The chimera NLRP3(1–217)-NLRP6(196–892) was also able to increase the percentage of cells with ASC specks when Nlrp3−/− macrophages were activated with nigericin (Fig. 6D) and in HEK293T cells in a K+ efflux–dependent manner (Fig. 6E).
Nigericin was able to induce oligomerization of chimeric
NLRP3(1–217)-NLRP6(196–892) when expressed in HEK293T cells without ASC,
and the formation of these oligomers was abolished when high
extracellular KCl was used (Fig. 6F). This suggests that the NLRP3 PYD-linker-FISNA sequence is a key region to induce a conformational change in NLRP3 after K+ efflux and favor NLRP3 activation.
Fig. 6. NLRP3 PYD-linker-FISNA sequence renders NLRP6 sensitive to K+ efflux.
(A) IL-1β release from Nlrp3−/−
immortalized macrophages (iMos) treated for 16 hours with doxycycline
(1 μg/ml) and LPS (100 ng/ml) to induce the expression of YFP-NLRP6,
YFP-NLRP3, or YFP-chimeric NLRP3/6 as indicated and then activated for
60 min with nigericin (10 μM) in the absence/presence of 40 mM KCl; n = 4 independent experiments; Mann-Whitney test. (B) Western blot for IL-1β, caspase-1, NLRP3/6, and β-actin, from cell lysates and supernatants of Nlrp3−/− iMos expressing YFP-NLRP3, YFP-NLRP3(1–217)-NLRP6(196–892), or YFP-NLRP6 treated as in (A); representative of n = 3 independent experiments. (C) IL-1β release [enzyme-linked immunosorbent assay (ELISA), left; Western blot, right] from Nlrp3−/−
iMos expressing the chimera YFP-NLRP3(1–217)-NLRP6(196–892) treated as
in (A) in the absence/presence of MCC950 (10 μM) or Ac-YVAD-AOM (YVAD;
100 μM). For the ELISA, n = 4 to 6 independent experiments; Mann-Whitney test. (D) Percentage of ASC specking in Nlrp3−/− iMos expressing the chimera YFP-NLRP3(1–217)-NLRP6(196–892) treated as in (A). n = 7 to 8 independent experiments; Mann-Whitney test. (E)
Percentage of ASC specking in HEK293T cells expressing ASC and the
chimera YFP-NLRP3(1–217)-NLRP6(196–892) treated for 30 min with
nigericin (10 μM) in the absence/presence of 140 mM KCl. n = 5 to 15 independent experiments; Kruskal-Wallis test. (F)
Representative fluorescent micrographs of HEK293T cells expressing the
chimera YFP-NLRP3(1–217)-NLRP6(196–892) (green) and stained for nuclei
(DAPI, blue), treated as in (E). Scale bar, 10 μm. Right: Quantification
of YFP-NLRP3(1–217)-NLRP6(196–892) cells with a punctuate staining. n = 4 independent experiments; Mann-Whitney test.
The NLRP3 N-terminal domain is important for activation in response to crystals
Particulate matter, such as monosodium urate (MSU) crystals, activates NLRP3 by a decrease of the intracellular K+ concentration (12, 35). However, the kinetics of NLRP3 activation induced by crystals is slower than when K+ efflux is driven by an ionophore or the activation of ion channels (36), probably due to the slow dilution of intracellular K+ concentration when MSU induces cell swelling (35). We found that MSU induced IL-1β release from Nlrp3−/−
macrophages expressing NLRP3 with the deletion Δ92–120 and much less
with the deletion Δ92–132 but not when the deletion Δ92–148 or Δ92–217
was expressed (Fig. 7A). High extracellular K+ was able to significantly reduce IL-1β release induced by NLRP3 Δ92–120 and Δ92–132 (Fig. 7A), supporting the data observed with nigericin. The NLRP3 mutant AMAA132 was also able to release IL-1β in a K+-dependent manner upon MSU activation similarly to the polybasic RMRR132 mutant (fig. S8E), further supporting that human KMKK132
NLRP3 sequence is dispensable for activation. Swapping NLRP6 N terminus
sequence with the NLRP3 sequence resulted in a chimeric receptor that,
when expressed in Nlrp3−/− macrophages, released IL-1β
after MSU treatment when the NLRP3 sequence included the PYD and the
linker sequence (1–132) or the PYD-linker-FISNA (1–217), dependent on K+ efflux (Fig. 7B).
These results support the idea that the activation of NLRP3 induced by
MSU crystals could also be facilitated by the linker (92–132) sequence.
Fig. 7. NLRP6 with the PYD-linker-FISNA NLRP3 sequence is activated by MSU crystals and imiquimod.
(A) IL-1β release from Nlrp3−/−
immortalized macrophages (iMos) expressing different truncations of
NLRP3 as indicated treated for 16 hours with doxycycline (1 μg/ml), LPS
(100 ng/ml), and MSU crystals (300 μg/ml) in the absence/presence of 40
mM KCl. n = 3 to 4 independent experiments; Mann-Whitney test. (B) IL-1β release from Nlrp3−/− iMos treated as in (A) but expressing different chimeric NLRP6/3 receptors as indicated. n = 3 to 4 independent experiments; Mann-Whitney test. (C) IL-1β release from Nlrp3−/−
iMos expressing different truncations of NLRP3 as indicated, treated
for 16 hours with doxycycline (1 μg/ml) and LPS (100 ng/ml) to induce
the expression of different NLRP3 truncations as indicated and then
activated for 1 hour with imiquimod (100 μM) in the absence/presence of
40 mM KCl. n = 3 to 4 independent experiments; Mann-Whitney test. (D) IL-1β release from Nlrp3−/− iMos treated as in (C) but expressing different chimeric NLRP6/3 receptors as indicated. n = 3 to 4 independent experiments; Mann-Whitney test. (E)
Quantification of HEK293T cells expressing YFP-NLRP3 treated for
different times with imiquimod (100 μM, for 0.5, 2, 4, or 6 hours) or
nigericin (10 μM, 30 min) in the absence/presence of 40 mM KCl. n = 3 to 4 independent experiments; Mann-Whitney test. (F)
Average BRET signal for YFP-NLRP3-Luc expressed in HEK293T cells before
and after stimulation with imiquimod (100 μM, 6 hours), in the absence
or presence of 40 mM KCl. n = 6 to 9 independent measurements; Kruskal-Wallis test.
The linker sequence is important for NLRP3 activation in response to specific K+-independent NLRP3 stimulus
NLRP3 can also be activated by imiquimod and derivatives in a K+ efflux–independent manner (37), and the mouse NLRP3 polybasic sequence KKKK130 of the linker domain has been reported as critical for this activation (4).
Here, we found that the human NLRP3 Δ92–120 deletion, but not the
Δ92–132, Δ92–148, or Δ92–217 deletion, was able to induce IL-1β release
after imiquimod treatment when expressed in Nlrp3−/− macrophages (Fig. 7C). Accordingly, IL-1β release induced by imiquimod in NLRP3 Δ92–120 deletion was insensible to K+ efflux (Fig. 7C), indicating that the predicted α helix of the linker sequence found in the 120–132 sequence could be important for K+-independent stimulation of NLRP3. However, the KMKK132 polybasic sequence appears not to be critical for NLRP3 activation in response to imiquimod, as Nlrp3−/− macrophages expressing AMAA132 NLRP3 mutants were able to release IL-1β in a K+-independent
manner after imiquimod activation (fig. S8F). Furthermore, including
NLRP3 PYD-linker (1–132) or NLRP3 PYD-linker-FISNA (1–217) of NLRP3 in
NLRP6 background resulted in a chimeric receptor that, when expressed in
Nlrp3−/− macrophages, released IL-1β after imiquimod treatment (Fig. 7D). Imiquimod was able to induce NLRP3 oligomerization (Fig. 7E) and decreased NLRP3 BRET signal (Fig. 7F) independently of K+
efflux. These data support the idea that imiquimod could also induce a
conformational change in NLRP3 for activation and that the NLRP3 linker
sequence (92–132), and particularly the N-tail of the α helix, is key to
activate NLRP3 in response to imiquimod in a K+ efflux–independent manner.
NLRP6 with the NLRP3 PYD-linker-FISNA sequence induces an inflammatory response in vivo
MSU crystals induce inflammation through NLRP3 receptor activation (38).
NLRP3-deficient mice present less peritoneal IL-1β and less peritoneal
granulocyte infiltration when challenged intraperitoneally with MSU
crystals compared to wild-type mice (fig. S9A), being this an
established model for NLRP3-driving peritonitis (38). We developed a model using Nlrp3−/− mice intraperitoneally reconstituted with immortalized Nlrp3−/−
mouse macrophages expressing either NLRP6, the chimera
NLRP3(1–217)-NLRP6(196–892), or NLRP3 and then challenged with
intraperitoneal MSU crystals (Fig. 8A).
These macrophages were present in the peritoneum by the time MSU would
be administered and after the final recovery of peritoneal exudates
(fig. S9B). The recovered recombinant peritoneal macrophages were
functional, activating the inflammasome (fig. S9C). MSU crystals induced
an increase of IL-1β and IL-18 concentration in the peritoneum of Nlrp3−/− mice reconstituted with Nlrp3−/− macrophages expressing NLRP3 and the chimera NLRP3(1–217)-NLRP6(196–892), but not when NLRP6 was expressed (Fig. 8A).
IL-6 concentration in the peritoneum of these mice was very low and
similar between the different macrophages reconstituted into the
peritoneum (fig. S9D). Mice reconstituted with macrophages with the
chimeric NLRP3(1–217)-NLRP6(196–892) and full-length NLRP3 receptor
induced an increase of IL-1–dependent chemokines CXCL1, CXCL10, and CCL2
in the peritoneum of mice intraperitoneally injected with MSU crystals (Fig. 8B) and with a parallel increase of the percentage of infiltrating granulocytes (Fig. 8C).
This in vivo model indicates that the NLRP3 PYD-linker-FISNA sequence
is key to initiate an inflammatory response to MSU crystals.
Fig. 8. NLRP6 with the NLRP3 PYD-linker-FISNA sequence induces an inflammatory response to MSU crystals.
(A) Protocol followed in the experimental in vivo assay (top). IL-1β and IL-18 in the peritoneal lavage of Nlrp3−/− after an intraperitoneal (i.p.) injection with Nlrp3−/−
iMos expressing YFP-NLRP6, the chimera YFP-NLRP3(1–217)-NLRP6(196–892),
or YFP-NLRP3 and then challenged with an intraperitoneal injection of
MSU crystals for 16 hours (bottom). n = 7 to 11 mice, each one represented by a dot; Mann-Whitney test. (B) CXCL1, CXCL10, and CCL2 chemokines in the peritoneal lavage of Nlrp3−/− mice treated as in (A). n = 8 to 11 mice, each one represented by a dot; Mann-Whitney test. (C) Quantification of CD11b+, Ly6G+, and F4/80− cells in the peritoneal lavage of Nlrp3−/− mice treated as in (A). n = 5 to 11 mice, each one represented by a dot; Mann-Whitney test.
DISCUSSION
The
NLRP3 inflammasome has been implicated in multiple diseases, and
therefore, its activation mechanism involves diverse signaling steps
that remain not fully understood. However, most of NLRP3 triggers share
the requirement to induce a reduction in the intracellular concentration
of K+ (10, 11).
In the present study, we identified that the inactive NLRP3 protein
structure changes to favor activation in response to low concentrations
of intracellular K+. The linker and FISNA domain, a
NLRP3-specific domain, both located between the N-terminal PYD and
central NACHT domains, are important to activate NLRP3 when
intracellular K+ decreases and when K+-independent
activators are used. The notion that the LRR domain of NLRP3 could be
responsible for a ligand-binding activation model has been recently
challenged by the identification of a minimal NLRP3 sequence lacking the
LRR that is activated in a K+ efflux–dependent manner similarly to the full-length NLRP3 (31).
Our results support this model as we identified the region between the
PYD and NACHT domains as critical for NLRP3 activation. The interaction
of mouse NLRP3 with negatively charged lipids in the dispersed
trans-Golgi network by a polybasic sequence present in the region
between the NLRP3 PYD and NACHT domain (KKKK130) at the end of exon 3 is important for its activation (4). However, while the dispersion of trans-Golgi network induced by the different NLRP3 activators is independent of K+
efflux, the activation of NLRP3 is blocked by increasing extracellular
concentrations of KCl and therefore dependent on the efflux of K+ (4, 12).
To avoid possible artefacts of the use of elevated concentrations of
extracellular KCl, we also found that NLRP3 is equally blocked by
increasing extracellular RbCl, being Rb+ mimicking K+ conductance in most K+ permeable channels and ionophores. Our data indicate that the decrease of intracellular K+
concentration changes the inactive structure of NLRP3, resulting in a
conformation favoring the functional oligomerization of the receptor
into active oligomers. This conformational change is independent of the
partially conserved polybasic region of mouse NLRP3 that binds to
PtdIns4P (KKKK130), as the human NLRP3 (KMKK132) mutant lacking this region (Δ92–132) or the AMAA132
mutant was still undergoing the same conformational change as measured
by a reduction of the BRET signal. Although in the absence of a
full-length NLRP3 structure in active conformation it is difficult to
interpret the change in BRET signal, it suggests either that a
conformational change occurs before NLRP3 binding to negatively charged
lipids on the dispersed trans-Golgi network or that an additional
polybasic sequence found at the beginning of the FISNA domain in the
human NLRP3 (RKKYRKYVRSR145) could be important for PtdIns4P
binding in the dispersed trans-Golgi network, allowing then a
conformational change of the receptor. Additional deletion of this
second polybasic region in NLRP3 (Δ92–148) completely prevents NLRP3
activation and probably recruitment to the dispersed trans-Golgi
network, as this second polybasic sequence was also important for mouse
NLRP3 binding to PtdIns4P (4).
After NLRP3 binding to the dispersed trans-Golgi network, the receptor
is trafficked to the centrosome for full inflammasome activation, where
it probably facilitates its interaction with the centrosome-localized
kinase NEK7 and the adaptor protein ASC (39). Consistent with our study, the interaction of NEK7 with NLRP3 occurs after K+ efflux (3). An incomplete NLRP3 structure bound to NEK7 has been solved (28);
this NLRP3 structure is partially closed and is considered a
semi-inactive NLRP3. We found that the NLRP3 nucleotide binding site in
this NLRP3/NEK7 structure is locked, impeding the exchange of adenosine
diphosphate (ADP) for ATP, which is also required for NLRP3 activation (30). Therefore, the opening of the NLRP3 inactive structure induced by K+
efflux could facilitate the entrance of ATP to the nucleotide binding
pocket and NLRP3 activation. During cell swelling, the decrease of
intracellular K+ is responsible for a conformational change in the inactive NLRP3 molecules existing in preassembled inactive complexes (24).
Our present study indicates, using an intramolecular BRET sensors, that
the NLRP3 conformational change during activation is dependent on the
presence of the FISNA domain that includes a second polybasic sequence
(RKKYRKYVRSR145) possibly important for human NLRP3 binding
to PtdIns4P and activation. Furthermore, the linker sequence between the
PYD and FISNA (92–132) may assist on the correct placement of the NLRP3PYD domain to form a seed for ASCPYD
nucleation, thus supporting active oligomer formation. Therefore, the
NLRP3 N-terminal sequence is important for NLRP3 inflammasome activation
in response to K+ efflux.
Imiquimod and derivatives are activators of the NLRP3 inflammasome (40), but their mechanism of activation is independent of K+ efflux (37) and dependent on the binding of NLRP3 to PtdIns4P in the dispersed trans-Golgi network (4).
Our study demonstrates that imiquimod is also able to induce a decrease
of NLRP3 BRET signal, suggesting a similar conformational change on
NLRP3 than when nigericin is applicated and probably favoring the
activation of the inflammasome. We found that the presence of the linker
sequence (92–132), between the PYD and FISNA domain, is important for
NLRP3 activation by imiquimod, indicating that both K+-dependent and K+-independent
activation of NLRP3 would share a similar conformational change
involving a specific N-terminal domain on the receptor, because both
types of triggers induce a reduction in the NLRP3 BRET signal. We also
previously reported a decrease of the NLRP3 BRET signal in active NLRP3
mutants associated to autoinflammatory syndromes (22).
The NLRP6 inflammasome is triggered by some Gram-positive bacteria cell wall polymers such as lipoteichoic acid (33); however, it cannot be activated by classical NLRP3 triggers that induce K+ efflux (34).
Therefore, NLRP3 is unique among other inflammasome sensors because it
is the only one able to respond to specific damage- and
homeostasis-associated molecular patterns. By introducing the N-terminal
fragment of NLRP3 in NLRP6, we generated a chimeric receptor that is
activated in response to nigericin, MSU crystals, and imiquimod.
Therefore, the N-terminal sequence of NLRP3 is key for its activation in
response to specific triggers, except for the PYD domain that is not
necessary for NLRP3 activation and oligomerization, but is crucial to
create functional NLRP3 inflammasome by allowing the recruitment of ASC (13). The NLRP3 inhibitor MCC950 specifically binds to the NACHT domain of NLRP3 and affects the active structure of NLRP3 (22, 32),
and here, we found that MCC950 is not blocking the chimeric NLRP3/NLRP6
receptor, further supporting the specificity of MCC950 over NLRP3 NACHT
domain.
Together, these results reveal that
the NLRP3 sequence between the PYD and NACHT is important to confer
responsiveness of NLRP3 to specific triggers. Therefore, the unique
ability of NLRP3 to activate in response to signals that decrease
intracellular K+ resides in the presence of a linker sequence
encoded by a specific NLRP3 exon 3 and the FISNA domain, which are
necessary to facilitate a structural change of the closed inactive NLRP3
protein. Then, the linker sequence is also important to activate NLRP3
in response to K+-independent triggers and facilitates an optimal orientation of the NLRP3PYD domain within the active NLRP3 oligomer to form a helical seed to interact with ASCPYD and form a fully active inflammasome.
MATERIALS AND METHODS
Plasmid construction
The
different constructs of human NLRP3 and NLRP6 were generated by
overlapping polymerase chain reaction (PCR) (UniProt #Q96P20 and #P59044
annotations for human NLRP3 and NLRP6, respectively) and cloned into
pcDNA3.1/V5-His TOPO (Life Technologies). Sequencing of all constructs
was performed to confirm correct modification and the absence of
unwanted mutations. All constructs were designed to contain YFP at the N
terminus for microscopy assays or double-tagged with YFP at the N
terminus and Renilla luciferase (Luc) at the C terminus to generate the various BRET sensors. NLRP3 containing Renilla
luciferase (Luc) at the C terminus was also constructed in
pcDNA3.1/V5-His TOPO, sequenced to confirm correct alignment between the
tag and the NLRP3 sequence, and used as control in all BRET assays.
Cells and transfections
HEK293T
cells (CRL-11268, American Type Culture Collection) were maintained in
Dulbecco’s modified Eagle’s medium (DMEM)/F-12 (1:1) (Lonza)
supplemented with 10% fetal calf serum (FCS) (Life Technologies), 2 mM
GlutaMAX (Life Technologies), and 1% penicillin-streptomycin (Life
Technologies). HEK293T cells stably expressing the rat P2X7 receptor
have previously been described (41)
and were cultured in F-12 media (Lonza) supplemented with 10% FCS.
Lipofectamine 2000 was used for the transfection of HEK293T cells
according to the manufacturer’s instructions. After 2 days of
transfection, stable selection of HEK293T clones expressing the
different NLRP3 constructs was initiated by supplementing cell culture
media with G418 (2 mg/ml; Acros Organics). After 4 weeks of culture in
G418, cell cloning was performed by serial dilution in 96-well plates in
the presence of G418 for a further 4 to 8 weeks. Positive clones were
expanded and tested for correct expression by Western blot and
fluorescence microscopy. HEK293T cells stably expressing NLRP3
constructs were maintained in DMEM/F-12 (1:1) supplemented with 10% FCS,
2 mM GlutaMAX, and 1% penicillin-streptomycin. All cells were routinely
tested for mycoplasma contamination with the Mycoplasma Detection Kit
(Roche).
Differentiation and stimulation of bone marrow–derived macrophages
Bone
marrow–derived macrophages (BMDMs) were obtained from wild-type mice by
differentiating bone marrow cells for 7 days in DMEM (Lonza)
supplemented with 25% of L929 medium, 15% FCS, penicillin-streptomycin
(100 U/ml), and 2 mM GlutaMAX as described elsewhere (42). Cells were primed for 4 hours with Escherichia coli LPS O55:B5 (1 μg/ml; InvivoGen), then washed in E-total buffer (147 mM NaCl, 10 mM Hepes, 13 mM glucose, 2 mM CaCl2, 1 mM MgCl2,
and 2 mM KCl, pH 7.4), and treated for 30 min with nigericin (10 μM;
Sigma-Aldrich), for 2 hours with valinomycin (50 μM; Sigma-Aldrich), or
for 2 hours with BB15C5 (50 μM; Sigma-Aldrich). In some experiments,
stimulations were performed in an E-total buffer with increased
concentration of KCl or changing KCl for RbCl, LiCl, or CsCl (as denoted
in the figure legends); in that case, the increase in the concentration
of KCl, RbCl, LiCl, or CsCl was accompanied with a reduction of the
NaCl concentration to maintain isotonic conditions.
Retroviral production, immortalized macrophage generation, and stimulation
For
doxycycline-inducible expression of NLRP constructs in immortalized
mouse macrophages, we used Tet-ON retroviral system (#631188, Clontech).
The different NLRP constructs (deletions or chimeras) were subcloned
into pRETROX Tre3G plasmid (Clontech) using Bam HI/Eco RI and
transfected using Lipofectamine 2000 into the packaging cell line
Gryphon Ampho cell line (Alelle Biotechnology, ABP-RVC-10001). Nlrp3−/− immortalized mouse macrophages stably expressing the Tet-On 3G transactivator (31)
were transduced with different NLRP constructs or empty vector encoding
retroviruses for 2 days. Then, positive macrophages were selected with
puromycin (6 μg/ml) and G418 (1.5 mg/ml). For experiments, immortalized
mouse macrophages were treated for 16 hours with doxycycline (1 μg/ml;
Sigma-Aldrich) and ultrapure LPS 0111:B4 (100 ng/ml; InvivoGen) and then
stimulated for 1 hour with nigericin (10 μM; Sigma-Aldrich) or TcdB (1
μg/ml; BML-G150-0050, Enzo), for 6 hours with imiquimod (100 μM;
tlrl-imqs, InvivoGen), or for 16 hours with MSU crystals (300 μg/ml;
ALX-400-047-M002, Enzo) or with transfected lipoteichoic acid (15 μg/ml;
tlrl-pslta, InvivoGen). The specific NLRP3 inhibitor MCC950 (10 μM;
CP-456773, Sigma-Aldrich) or the caspase-1 inhibitor IV Ac-YVAD-AOM (100
μM; 400015, Calbiochem) was added 30 min before and during the
different stimulations.
In vivo MSU model
All
experimental protocols for animal handling were refined and approved by
the Ethical Committee for Animal Research of the University of Murcia
(reference 542/2019). C57BL/6J mice (wild type) were obtained from
Charles River and inbred at the specific pathogen–free animal house of
the IMIB-Arrixaca up to F3 generation before getting new founders, and
NLRP3-deficient mice (Nlrp3−/−) in C57BL/6J background were already described (38).
Mice were bred in specific pathogen–free conditions with a 12:12-hour
light-dark cycle and used in accordance with the Spanish national (RD
1201/2005 and Law 32/2007) and EU (86/609/EEC and 2010/63/EU)
legislation. Nlrp3−/− immortalized mouse macrophages
expressing different NLRs or chimeric NLRs were treated in vitro for 16
hours with doxycycline (1 μg/ml). Besides, mice were given doxycycline
(2 mg/ml) and sucrose (5%, Sigma-Aldrich) in the drinking water 1 day
before experimental initiation. Mice between 8 and 10 weeks of age
receive an intraperitoneal injection of 2 × 106 immortalized
macrophages, and in some experiments, the macrophages were stained for
15 min with carboxyfluorescein succinimidyl ester (10 μM; Thermo Fisher
Scientific) before injection. After 3 hours of macrophages homing in the
recipients, the animals received an intraperitoneal injection of MSU
crystals (40 mg/kg; ALX-400-047-M002, Enzo). After 16 hours, animals
were euthanized with CO2 inhalation and peritoneal lavages
were collected after exposing the abdominal wall by opening the skin,
and 4 ml of sterile saline solution was injected into the peritoneal
cavity via a 25-gauge needle. The abdomen was gently massaged for 1 min,
and the peritoneal fluid was recovered through the needle and
centrifuged at 433g for 10 min. The cellular pellet was
immediately used for flow cytometry, and the supernatants were stored at
−80°C until further enzyme-linked immunosorbent assay (ELISA) analysis.
Flow cytometry
Cells (5 × 105)
from mouse peritoneal lavage were incubated with anti-CD16/32 (clone
93, 14-0161-85, eBioscience; 1:200) and then stained with
anti-F4/80–Alexa Fluor 488 (clone BM8, 123119, BioLegend; 1:200),
anti-CD11b–APC (allophycyanin) (clone M1/70, 101211, BioLegend; 1:200),
and anti-Ly6G–PE (phycoerythrin) (clone 1A8, 127607, BioLegend; 1:200).
Granulocytes were identified as CD11b+F4/80−Ly6G+ (fig. S10A). In some experiments, the presence of injected immortalized macrophages was monitored by F4/80+CSFE+
cells (fig. S10B); for that, the anti-F4/80–APC antibody was used
(clone BM8, 123115, BioLegend; 1:200). Samples were analyzed in
FACSCanto (BD Biosciences) by gating for singlets based on forward and
side scatter parameters (fig. S10, A and B). The data were analyzed by
FCS Express 5 software (De Novo Software).
ELISA and multiplex assay
Cell-free
peritoneal lavage and macrophage supernatants were tested by ELISA for
mouse IL-1β or IL-6 following the manufacturer’s instructions (R&D
Systems and eBioscience, respectively) and read in a Synergy Mx (BioTek)
plate reader. Multiplexing for CXCL1, CXCL10, IL-18, IL-6, and CCL2
from peritoneal lavages was performed using the ProcartaPlex Multiplex
Immunoassay (Invitrogen) following the manufacturer’s indications and
analyzed in a Bio-Plex analyzer (Bio-Rad).
Western blot
Cells
were lysed for 30 min on ice in cold lysis buffer [50 mM tris-HCl (pH
8.0), 150 mM NaCl, 2% Triton X-100, supplemented with protease inhibitor
mixture (100 μl/ml) from Sigma-Aldrich] and then were centrifuged at
16,000g for 15 min at 4°C. Proteins in cell culture supernatants
were concentrated using a 10-kDa cutoff column (Microcon, Merck
Millipore) by centrifugation at 11,200g for 30 min at 4°C. Cells
lysates and concentrated supernatants were resolved in 4 to 12% precast
Criterion polyacrylamide gels (Bio-Rad) and transferred to
nitrocellulose membranes (Bio-Rad) by electroblotting. Membranes were
probed with anti-GSDMD rabbit monoclonal (EPR19828, ab209845, Abcam;
1:5000), anti–IL-1β rabbit polyclonal (H-153; sc-7884; 1:1000),
anti–caspase-1 rabbit polyclonal (sc-514, Santa Cruz Biotechnology;
1:1000), horseradish peroxidase (HRP) anti–β-actin (C4; sc-47778HRP,
Santa Cruz Biotechnology; 1:10,000), and anti–green fluorescent protein
rabbit polyclonal antibodies (ab6556, Abcam; 1:2500). HRP-conjugated
secondary antibodies were from GE Healthcare. Full uncropped Western
blots are presented in fig. S11.
Quantitative reverse transcription PCR analysis
Total
RNA purification was performed using the RNeasy Kit (Qiagen) according
to the manufacturer’s recommendations and quantified on NanoDrop 2000
(Thermo Fisher Scientific). Detailed methods used for quantitative
reverse transcription PCR have been described previously (42).
Briefly, reverse transcription was realized using the iScript cDNA
Synthesis Kit (Bio-Rad). Quantitative PCR was performed in the iQ 5
Real-Time PCR Detection System (Bio-Rad) with an SYBR Green mix
(Takara), and primers used were obtained from Sigma-Aldrich (KiCqStart
Primers) The presented relative gene expression levels were calculated
using the 2−ΔCt method normalizing to Hprt1 expression as endogenous control.
Bioluminescence resonance energy transfer
HEK293T cells expressing the different NLRP3 BRET sensors (wild type, mutations, and deletions) were plated on a poly-l-lysine–coated
white opaque 96-well plate; after adhesion, cells were incubated with
different treatments in E-total (with different ionic composition, as
stated in the figure legends) or vehicle and BRET readings were
performed 5 min after the addition of 5 μM coelenterazine-H substrate.
BRET signals were detected with two filter settings [Renilla
luciferase (Luc) filter (485 ± 20 nm) and YFP filter (530 ± 25 nm)] at
37°C using a Synergy Mx plate reader (BioTek) as described before (25).
In some experiments, BRET signal was recorded every 35 s before and
after nigericin, valinomycin, BB15C5, or ATP (this one over HEK293T
cells expressing P2X7 receptor) automatic injection for a total of 15
min. For experiments measuring basal BRET signal, a stable signal for
5-min kinetic was recorded and averaged. For some experiments, nigericin
or imiquimod was added to the plate and placed in the incubator for 60
min or 6 hours, respectively, and then coelenterazine-H was added; after
5 min of signal stabilization, BRET was recorded for 20 min. Titration
was performed by transfection of different amounts of the plasmids used
in this study (pcDNA empty plasmid was used to have equal amounts of
total DNA in all the transfections). Expression of the different sensors
was monitored by reading YFP fluorescence in the plate reader or by
assessing individual cell relative fluorescence by fluorescence
microscopy. Titration of the sensor will determine whether the recorded
BRET is intra- or intermolecular, because intramolecular energy transfer
results in a stable BRET signal as the BRET sensor concentration
increases and intermolecular BRET will result in a proportional
increases of the BRET signal as the sensor concentration increases (25).
The BRET ratio was defined as the difference of the emission ratio 530
nm/485 nm of the BRET sensor minus this ratio of the Luc only–tagged
NLRP3. Results were expressed in milliBRET units (mBU).
Fluorescence microcopy
Following procedures previously described in (22), we seeded 5 × 104 HEK293T cells or Nlrp3−/− immortalized macrophages expressing the different NLRP constructs tagged with YFP at the N terminus in poly-l-lysine–coated
coverslips (Corning). Cells were treated and stimulated as indicated in
the figure legends, washed twice with phosphate-buffered saline (PBS),
fixed for 15 min at room temperature with 4% paraformaldehyde, and then
washed three times with PBS. For ASC immunofluorescence, procedures have
been previously described in (22);
in particular, cells were blocked with 2% bovine serum albumin and
permeabilized with 0.2% Triton X-100 (Sigma-Aldrich) for 20 min at room
temperature. Then, cells were stained for 1.5 hours at room temperature
with the primary monoclonal mouse antibody anti-ASC (HASC-71, BioLegend;
1:1000). Cells were washed and then incubated for 1 hour at room
temperature with anti-mouse immunoglobulin G (IgG)
fluorescence-conjugated secondary antibody [Alexa 647 donkey anti-mouse
IgG (H+L), Life Technologies; 1:200]. Cells were washed and nuclei were
stained for 10 min with 4′,6-diamidino-2-phenylindole (1:10,000;
Sigma-Aldrich), and coverslips were mounted on slides with mounting
medium (S3023, Dako). Images were acquired using the same equipment that
in previous studies (22),
and we used a Nikon Eclipse Ti microscope equipped with a 20× S Plan
Fluor objective (numerical aperture, 0.45), a 60× Plan Apo Vc objective
(numerical aperture, 1.40), and a digital Sight DS-QiMc camera (Nikon)
with a Z optical with spacing of 0.4 μm and 387-/447-nm, 472-/520-nm,
543-/593-nm, and 650-/668-nm filter sets (Semrock) and the NIS-Elements
AR software (Nikon). Images were analyzed with ImageJ (U.S. National
Institutes of Health).
Measurement of intracellular K+
Twelve-well plates with 106
BMDMs per well were stimulated for 4 hours with LPS (1 μg/ml) at 37°C,
then washed twice with E-total buffer, and stimulated for 30 min at 37°C
with different concentrations of nigericin as indicated in the figure
legends or for 2 hours with valinomycin (50 μM) or BB15C5 (50 μM). Then,
cells were briefly and quickly washed with nuclease-free water to avoid
an osmotic shock and immediately after cells were scraped in 200 μl per
well of nuclease-free water followed by three freeze-thaw cycles.
Lysates were centrifuged at 16,000g for 10 min at 4°C, and the supernatants were stored at −80°C until K+ concentration was quantified by indirect potentiometry using Cobas 6000 with ISE module (Roche).
Bioinformatic analysis and modeling
NLRP3
sequence annotation and numbering was used from UniProt database
(Q96P20). Multiple protein sequence alignment was performed using
Clustal omega (43). FISNA domain signature was obtained by SMART database (44), and secondary FISNA domain prediction was obtained by Jpred 4 (45).
The structure of human NLRP3 in complex with NEK7 and ADP as ligand in
the NACHT domain is taken from the Protein Data Bank (PDB) (46) structure with code 6NPY (28).
Several loop regions are missing in this structure, as well as the PYD
domain, the linker, and most of the FISNA domain. We used the structure
of NAIP2/NLRC4 inflammasome complex (3JBL) to infer the relocation of
NACHT and LRR domains in the active oligomeric conformation of NLRP3.
Then, we used the superposition between the monomers of NLRP3 and NLRC4
to identify the fragments of NLRP3 that move when opening the
conformation of NLRP3. The hinge is found in the loop region 417–441
(inside HD1 motif) between α helices 9 and 10 of the NACHT domain (fig.
S2A). We split the closed structure of NLRP3 around the hinge and
superposed with MatchMaker (47)
the two separated fragments, including the interaction with NEK7 and
ADP as ligand (fig. S12A). We constructed the scaffold of the open
conformation of NLRP3 by merging the two separated and superposed
fragments. We then used the scaffold of NLRP3 in open conformation as
template to model the sequence of NLRP3 starting at position 135 (i.e.,
except for the N-tail fragment, composed by the linker and the PYD
domain). We used SABLE (48)
to predict the secondary structure of the missing loops in the
structure of NLRP3 and NEK7. Helices are predicted at NLRP3 positions
113–128 (in the linker between FISNA and PYD domains), 182–190 (in the
FISNA domain), and 180–193 (in the structure of NEK7). A strand is
formed in NLRP3 between residues 174–179 of the FISNA domain, forming a
sheet with strand 364–369 with main-chain hydrogen-bonding interactions
in anti-parallel, stabilized by an α helix (fig. S12B). The rest of the
loops, in particular fragments 581–618 and 655–684 of NLRP3 and 180–193
of NEK7, were modeled with ModLoop (49), restricting the predicted helices (fig. S12B). We then used the structures of NAIP2/NLRC4 complex (3JBL) and the ASCPYD assembly (3J63) to construct a model of the NLRP3 complex of inflammasome that includes the oligomerization of ASCPYD.
First, 11 models of monomeric NLRP3 in open conformation, interacting
with NEK7 and ADP, without the PYD domain and the linker fragment that
joints PYD and FISNA domains, were superposed with each monomer of
NAIP2/NLRC4 in the structure 3JBL. The monomers of NLRP3 and NEK7 are
merged into a single PDB file, forming a complex of 22 monomers. Second,
we manually placed the complex of ASCPYD assembly near the N-tail of the monomeric models of NLRP3, forming a starting fiber of ASCPYD
and preserving the symmetry of the complex (fig. S13A). Third, we used
the PYD domain of NLRP3 from 3QF2. Then, taking into account that the
complex of ASCPYD is formed by layers of six monomers, we built the model of the full sequence of NLRP3 (fig. S13A) with MODELLER (50) upon an artificial template construction where the closest 11 monomers of ASCPYD to NLRP3NACHT were substituted by the PYD domain of NLRP3 as follows: (i) For each pair of close monomers of NLRP3, we associated an ASCPYD chain of the closest layer and another from the next layer (fig. S13B); (ii) we selected the initial 11 chains of ASCPYD
and superpose 11 models of the PYD domain of NLRP3 in each chain; and
(iii) we merged in the same chain one PYD domain with the rest of the
structure of NLRP3 (C-tail fragment composed by domains FISNA, NACHT,
and LRR) associated with it. Last, we used another structure of 3J63 to
superpose 13 additional chains of ASC. We used MODELLER to model the
whole complex, formed by 13 ASCPYD chains, 11 full chains of
NLRP3, and 11 chains of a partial structure of NEK7 (only the domain
with structure in 6NPY). We forced the symmetry between the chains of
NEK7 and the chain fragments composed by FISNA-NACHT-LRR domains of
NLRP3. The structure of the linker was modeled as a loop, but forcing an
α helix in position 113–128 (fig. S13C). The different symmetry between
NLRP3/NEK7 (with a C11 rotation axis) and the fiber of ASCPYD (with a C6 rotation axis) left the starting position of a chain of ASCPYD as a seed to continue forming the ASCPYD fiber (fig. S13D).
Statistical analysis
Statistical
analyses were performed using GraphPad Prism 7 (GraphPad Software,
Inc). Outliers were identified using the ROUT method and removed from
statistics. For two-group comparisons, Mann-Whitney test was used; when
comparing three or more groups, Kruskal-Wallis test was used. All data
are shown as mean values, and error bars represent standard error from
the number of independent assays indicated in the figure legend and
plotted in histograms as dots. P value is indicated as *P < 0.05; **P < 0.01; ***P < 0.001; P > 0.05 not significant (ns).
Acknowledgments
We thank K. A. Fitzgerald for immortalized NLRP3-deficient macrophages and I. Couillin for Nlrp3−/−
mice. We also thank A. I. Gómez (IMIB-Arrixaca, Murcia, Spain) for
technical assistance with molecular and cellular biology, F. Noguera and
M. Martínez (IMIB-Arrixaca, Murcia, Spain) for running the Hitachi ion
detection system, and the members of the Pelegrin’s laboratory, in
particular M. Mateo-Tórtola, for comments and suggestions throughout the
development of this project. We also want to acknowledge the support of
the SPF-animal house from IMIB-Arrixaca. Funding: I.H.-B. would
like to acknowledge the funding by the Slovenian Research Agency
(project grant J3-1746 and core funding P4-0176), and B.O. would like to
acknowledge the funding by the Ministerio de Economía, Industria y
Competitividad and ERDF (BIO2017-85329-R). This work was supported by
grants to A.T.-A. from the internal support program of the Medical
Faculty, University of Tübingen, Fortüne-Antrag Nr. 2615-0-0 and to P.P.
from FEDER/Ministerio de Ciencia, Innovación y Universidades—Agencia
Estatal de Investigación (grant SAF2017-88276-R), Fundación Séneca
(grants 20859/PI/18, 21081/PDC/19, and 0003/COVI/20), and the European
Research Council (ERC-2013-CoG grant 614578 and ERC-2019-PoC grant
899636). Author contributions: A.T.-A. and D.A.-B. equally
performed most of the experimental work; C.A.-V. performed the in vivo
model; M.C.B. performed part of the molecular biology, transfections,
and stable cell line generation; I.H.-B. set up retroviral expression
system in immortalized macrophage lines; A.T.-A. established the
different immortalized macrophage cell lines and conceived and performed
analysis of the NLRP3 exon 3 linker sequence; B.O. and D.A.-B.
performed bioinformatic structural analysis; A.T.-A., D.A.-B., C.A.-V.,
and P.P. analyzed the data, interpreted results, and conceived the
experiments; P.P. prepared the figures with the help of A.T.-A.,
D.A.-B., C.A.-V., and B.O.; A.T.-A., D.A.-B., B.O., I.H.-B., and C.A.-V.
contributed to paper writing; P.P. conceived the project, provided
funding, wrote the paper, and supervised this study. Competing interests:
P.P. declares that he is an inventor in a patent filed on 12 March 2020
by the Fundación para la Formación e Investigación Sanitaria de la
Región de Murcia (PCT/EP2020/056729) for a method to identify
NLRP3-immunocompromised sepsis patients. The remaining authors declare
no competing interests. Data and materials availability: All data
needed to evaluate the conclusions of the paper are present in the
paper and/or the Supplementary Materials. No materials used in this
study are subject to a material transfer agreement. All materials are
available upon reasonable request to the corresponding author.
Supplementary Materials
This PDF file includes:
Figs. S1 to S13
Other Supplementary Material for this manuscript includes the following:
Source data for Figs. 1 to 8
REFERENCES AND NOTES
1
K. V. Swanson, M. Deng, J. P.-Y. Ting, The NLRP3 inflammasome: Molecular activation and regulation to therapeutics. Nat. Rev. Immunol. 19, 477–489 (2019).
2
M.
S. J. Mangan, E. J. Olhava, W. R. Roush, H. M. Seidel, G. D. Glick, E.
Latz, Targeting the NLRP3 inflammasome in inflammatory diseases. Nat. Rev. Drug Discov. 17, 588–606 (2018).
3
Y.
He, M. Y. Zeng, D. Yang, B. Motro, G. Núñez, NEK7 is an essential
mediator of NLRP3 activation downstream of potassium efflux. Nature 530, 354–357 (2016).
4
J. Chen, Z. J. Chen, PtdIns4P on dispersed trans-Golgi network mediates NLRP3 inflammasome activation. Nature 564, 71–76 (2018).
5
R.
Zhou, A. Tardivel, B. Thorens, I. Choi, J. Tschopp,
Thioredoxin-interacting protein links oxidative stress to inflammasome
activation. Nat. Immunol. 11, 136–140 (2009).
6
N.
Subramanian, K. Natarajan, M. R. Clatworthy, Z. Wang, R. N. Germain,
The adaptor MAVS promotes NLRP3 mitochondrial localization and
inflammasome activation. Cell 153, 348–361 (2013).
7
X.
Li, S. Thome, X. Ma, M. Amrute-Nayak, A. Finigan, L. Kitt, L. Masters,
J. R. James, Y. Shi, G. Meng, Z. Mallat, MARK4 regulates NLRP3
positioning and inflammasome activation through a microtubule-dependent
mechanism. Nat. Commun. 8, 15986 (2017).
8
P.
Samir, S. Kesavardhana, D. M. Patmore, S. Gingras, R. K. S. Malireddi,
R. Karki, C. S. Guy, B. Briard, D. E. Place, A. Bhattacharya, B. R.
Sharma, A. Nourse, S. V. King, A. Pitre, A. R. Burton, S. Pelletier, R.
J. Gilbertson, T.-D. Kanneganti, DDX3X acts as a live-or-die checkpoint
in stressed cells by regulating NLRP3 inflammasome. Nature 573, 590–594 (2019).
9
Y.
Duan, L. Zhang, D. Angosto-Bazarra, P. Pelegrín, G. Núñez, Y. He, RACK1
mediates NLRP3 inflammasome activation by promoting NLRP3 active
conformation and inflammasome assembly. Cell Rep. 33, 108405 (2020).
10
T. Próchnicki, M. S. Mangan, E. Latz, Recent insights into the molecular mechanisms of the NLRP3 inflammasome activation. F1000Res. 5, 1469 (2016).
11
I. Hafner-Bratkovič, P. Pelegrín, Ion homeostasis and ion channels in NLRP3 inflammasome activation and regulation. Curr. Opin. Immunol. 52, 8–17 (2018).
12
R. Muñoz-Planillo, P. Kuffa, G. Martínez-Colón, B. Smith, T. Rajendiran, G. Núñez, K+ efflux is the common trigger of NLRP3 inflammasome activation by bacterial toxins and particulate matter. Immunity 38, 1142–1153 (2013).
13
A.
Lu, V. G. Magupalli, J. Ruan, Q. Yin, M. K. Atianand, M. R. Vos, G. F.
Schröder, K. A. Fitzgerald, H. Wu, E. H. Egelman, Unified polymerization
mechanism for the assembly of ASC-dependent inflammasomes. Cell 156, 1193–1206 (2014).
14
F.
I. Schmidt, A. Lu, J. W. Chen, J. Ruan, C. Tang, H. Wu, H. L. Ploegh, A
single domain antibody fragment that recognizes the adaptor ASC defines
the role of ASC domains in inflammasome assembly. J. Exp. Med. 213, 771–790 (2016).
15
M.
S. Dick, L. Sborgi, S. Rühl, S. Hiller, P. Broz, ASC filament formation
serves as a signal amplification mechanism for inflammasomes. Nat. Commun. 7, 11929 (2016).
16
D.
Boucher, M. Monteleone, R. C. Coll, K. W. Chen, C. M. Ross, J. L. Teo,
G. A. Gomez, C. L. Holley, D. Bierschenk, K. J. Stacey, A. S. Yap, J. S.
Bezbradica, K. Schroder, Caspase-1 self-cleavage is an intrinsic
mechanism to terminate inflammasome activity. J. Exp. Med. 215, 827–840 (2018).
17
P. Broz, P. Pelegrín, F. Shao, The gasdermins, a protein family executing cell death and inflammation. Nat. Rev. Immunol. 20, 143–157 (2020).
18
A.
Baroja-Mazo, F. Martín-Sánchez, A. I. Gomez, C. M. Martínez, J.
Amores-Iniesta, V. Compan, M. Barberà-Cremades, J. Yagüe, E. Ruiz-Ortiz,
J. Antón, S. Buján, I. Couillin, D. Brough, J. I. Arostegui, P.
Pelegrín, F. Martin-Sanchez, A. I. Gomez, C. M. Martínez, J.
Amores-Iniesta, V. Compan, M. Barberà-Cremades, J. Yagüe, E. Ruiz-Ortiz,
J. Antón, S. Buján, I. Couillin, D. Brough, J. I. Aróstegui, P.
Pelegrín, The NLRP3 inflammasome is released as a particulate danger
signal that amplifies the inflammatory response. Nat. Immunol. 15, 738–748 (2014).
19
B.
S. Franklin, L. Bossaller, D. De Nardo, J. M. Ratter, A. Stutz, G.
Engels, C. Brenker, M. Nordhoff, S. R. Mirandola, A. Al-Amoudi, M. S.
Mangan, S. Zimmer, B. G. Monks, M. Fricke, R. E. Schmidt, T. Espevik, B.
Jones, A. G. Jarnicki, P. M. Hansbro, P. Busto, A. Marshak-Rothstein,
S. Hornemann, A. Aguzzi, W. Kastenmüller, E. Latz, The adaptor ASC has
extracellular and “prionoid” activities that propagate inflammation. Nat. Immunol. 15, 727–737 (2014).
20
C.
Venegas, S. Kumar, B. S. Franklin, T. Dierkes, R. Brinkschulte, D.
Tejera, A. Vieira-Saecker, S. Schwartz, F. Santarelli, M. P. Kummer, A.
Griep, E. Gelpi, M. Beilharz, D. Riedel, D. T. Golenbock, M. Geyer, J.
Walter, E. Latz, M. T. Heneka, Microglia-derived ASC specks cross-seed
amyloid-β in Alzheimer’s disease. Nature 552, 355–361 (2017).
21
V.
Petrilli, S. Papin, C. Dostert, A. Mayor, F. Martinon, J. Tschopp,
Activation of the NALP3 inflammasome is triggered by low intracellular
potassium concentration. Cell Death Differ. 14, 1583–1589 (2007).
22
A.
Tapia-Abellán, D. Angosto-Bazarra, H. Martínez-Banaclocha, C. de
Torre-Minguela, J. P. Cerón-Carrasco, H. Pérez-Sánchez, J. I. Arostegui,
P. Pelegrin, MCC950 closes the active conformation of NLRP3 to an
inactive state. Nat. Chem. Biol. 15, 560–564 (2019).
23
E. Bartok, M. Kampes, V. Hornung, in Methods in Molecular Biology, F. Di Virgilio, P. Pelegrin, Eds. (Springer Basel, 2016), vol. 1417, pp. 97–113.
24
V.
Compan, A. Baroja-Mazo, G. López-Castejón, A. I. Gomez, C. M. Martínez,
D. Angosto, M. T. Montero, A. S. Herranz, E. Bazán, D. Reimers, V.
Mulero, P. Pelegrín, Cell volume regulation modulates NLRP3 inflammasome
activation. Immunity 37, 487–500 (2012).
25
F. Martín-Sánchez, V. Compan, P. Pelegrín, in Methods in Molecular Biology, F. Di Virgilio, P. Pelegrin, Eds. (Springer Basel, 2016), vol. 1417, pp. 159–168.
26
F.
Martín-Sánchez, C. Diamond, M. Zeitler, A. I. Gomez, A. Baroja-Mazo, J.
Bagnall, D. Spiller, M. White, M. J. D. Daniels, A. Mortellaro, M.
Peñalver, P. Paszek, J. P. Steringer, W. Nickel, D. Brough, P. Pelegrín,
Inflammasome-dependent IL-1β release depends upon membrane
permeabilisation. Cell Death Differ. 23, 1219–1231 (2016).
27
M.
Cocco, C. Pellegrini, H. Martínez-Banaclocha, M. Giorgis, E. Marini, A.
Costale, G. Miglio, M. Fornai, L. Antonioli, G. López-Castejón, A.
Tapia-Abellán, D. Angosto, I. Hafner-Bratkovič, L. Regazzoni, C.
Blandizzi, P. Pelegrín, M. Bertinaria, Development of an acrylate
derivative targeting the NLRP3 inflammasome for the treatment of
inflammatory bowel disease. J. Med. Chem. 60, 3656–3671 (2017).
28
H.
Sharif, L. Wang, W. L. Wang, V. G. Magupalli, L. Andreeva, Q. Qiao, A.
V. Hauenstein, Z. Wu, G. Núñez, Y. Mao, H. Wu, Structural mechanism for
NEK7-licensed activation of NLRP3 inflammasome. Nature 570, 338–343 (2019).
29
L.
Zhang, S. Chen, J. Ruan, J. Wu, A. B. Tong, Q. Yin, Y. Li, L. David, A.
Lu, W. L. Wang, C. Marks, Q. Ouyang, X. Zhang, Y. Mao, H. Wu, Cryo-EM
structure of the activated NAIP2-NLRC4 inflammasome reveals nucleated
polymerization. Science 350, 404–409 (2015).
30
J.
A. Duncan, D. T. Bergstralh, Y. Wang, S. B. Willingham, Z. Ye, A. G.
Zimmermann, J. P.-Y. Ting, Cryopyrin/NALP3 binds ATP/dATP, is an ATPase,
and requires ATP binding to mediate inflammatory signaling. Proc. Natl. Acad. Sci. U.S.A. 104, 8041–8046 (2007).
31
I.
Hafner-Bratkovič, P. Sušjan, D. Lainšček, A. Tapia-Abellán, K. Cerović,
L. Kadunc, D. Angosto-Bazarra, P. Pelegrίn, R. Jerala, NLRP3 lacking
the leucine-rich repeat domain can be fully activated via the canonical
inflammasome pathway. Nat. Commun. 9, 5182 (2018).
32
R.
C. Coll, J. R. Hill, C. J. Day, A. Zamoshnikova, D. Boucher, N. L.
Massey, J. L. Chitty, J. A. Fraser, M. P. Jennings, A. A. B. Robertson,
K. Schroder, MCC950 directly targets the NLRP3 ATP-hydrolysis motif for
inflammasome inhibition. Nat. Chem. Biol. 15, 556–559 (2019).
33
H.
Hara, S. S. Seregin, D. Yang, K. Fukase, M. Chamaillard, E. S. Alnemri,
N. Inohara, G. Y. Chen, G. Núñez, The NLRP6 inflammasome recognizes
lipoteichoic acid and regulates Gram-positive pathogen infection. Cell 175, 1651–1664.e14 (2018).
34
P.
K. Anand, R. K. Malireddi, J. R. Lukens, P. Vogel, J. Bertin, M.
Lamkanfi, T.-D. Kanneganti, NLRP6 negatively regulates innate immunity
and host defence against bacterial pathogens. Nature 488, 389–393 (2012).
35
V.
Rabolli, L. Wallemme, S. Lo Re, F. Uwambayinema, M. Palmai-Pallag, L.
Thomassen, D. Tyteca, J.-N. N. Octave, E. Marbaix, D. Lison, O. Devuyst,
F. Huaux, Critical role of aquaporins in interleukin 1β (IL-1β)-induced
inflammation. J. Biol. Chem. 289, 13937–13947 (2014).
36
A.
Di, S. Xiong, Z. Ye, R. K. S. Malireddi, S. Kometani, M. Zhong, M.
Mittal, Z. Hong, T.-D. Kanneganti, J. Rehman, A. B. Malik, The TWIK2
potassium efflux channel in macrophages mediates NLRP3
inflammasome-induced inflammation. Immunity 49, 56–65.e4 (2018).
37
C.
J. Groß, R. Mishra, K. S. Schneider, G. Médard, J. Wettmarshausen, D.
C. Dittlein, H. Shi, O. Gorka, P.-A. Koenig, S. Fromm, G. Magnani, T.
Ćiković, L. Hartjes, J. Smollich, A. A. B. Robertson, M. A. Cooper, M.
Schmidt-Supprian, M. Schuster, K. Schroder, P. Broz, C. Traidl-Hoffmann,
B. Beutler, B. Kuster, J. Ruland, S. Schneider, F. Perocchi, O. Groß, K+ efflux-independent NLRP3 inflammasome activation by small molecules targeting mitochondria. Immunity 45, 761–773 (2016).
38
A.
Mayor, A. Tardivel, F. Martinon, V. Pe, V. Pétrilli, A. Mayor, A.
Tardivel, J. Tschopp, Gout-associated uric acid crystals activate the
NALP3 inflammasome. Nature 440, 237–241 (2006).
39
V.
G. Magupalli, R. Negro, Y. Tian, A. V. Hauenstein, G. Di Caprio, W.
Skillern, Q. Deng, P. Orning, H. B. Alam, Z. Maliga, H. Sharif, J. J.
Hu, C. L. Evavold, J. C. Kagan, F. I. Schmidt, K. A. Fitzgerald, T.
Kirchhausen, Y. Li, H. Wu, HDAC6 mediates an aggresome-like mechanism
for NLRP3 and pyrin inflammasome activation. Science 369, eaas8995 (2020).
40
T.-D.
Kanneganti, N. Özören, M. Body-Malapel, A. Amer, J.-H. Park, L.
Franchi, J. Whitfield, W. Barchet, M. Colonna, P. Vandenabeele, J.
Bertin, A. Coyle, E. P. Grant, S. Akira, G. Núñez, Bacterial RNA and
small antiviral compounds activate caspase-1 through cryopyrin/Nalp3. Nature 440, 233–236 (2006).
41
M.
T. Young, P. Pelegrin, P. Pelegrin, A. Surprenant, P. Pelegrin, A.
Surprenant, Identification of Thr283 as a key determinant of P2X7
receptor function. Br. J. Pharmacol. 149, 261–268 (2006).
42
J.
J. Martínez-García, H. Martínez-Banaclocha, D. Angosto-Bazarra, C. de
Torre-Minguela, A. Baroja-Mazo, C. Alarcón-Vila, L. Martínez-Alarcón, J.
Amores-Iniesta, F. Martín-Sánchez, G. A. G. A. Ercole, C. M. C. M.
Martínez, A. González-Lisorge, J. Fernández-Pacheco, P. Martínez-Gil, S.
Adriouch, F. Koch-Nolte, J. Luján, F. Acosta-Villegas, P. Parrilla, C.
García-Palenciano, P. Pelegrin, P2X7 receptor induces mitochondrial
failure in monocytes and compromises NLRP3 inflammasome activation
during sepsis. Nat. Commun. 10, 2711 (2019).
43
F.
Sievers, A. Wilm, D. Dineen, T. J. Gibson, K. Karplus, W. Li, R. Lopez,
H. McWilliam, M. Remmert, J. Söding, J. D. Thompson, D. G. Higgins,
Fast, scalable generation of high-quality protein multiple sequence
alignments using Clustal Omega. Mol. Syst. Biol. 7, 539 (2011).
44
I. Letunic, P. Bork, 20 years of the SMART protein domain annotation resource. Nucleic Acids Res. 46, D493–D496 (2018).
45
A. Drozdetskiy, C. Cole, J. Procter, G. J. Barton, JPred4: A protein secondary structure prediction server. Nucleic Acids Res. 43, W389–W394 (2015).
46
S.
K. Burley, H. M. Berman, C. Bhikadiya, C. Bi, L. Chen, L. Di Costanzo,
C. Christie, K. Dalenberg, J. M. Duarte, S. Dutta, Z. Feng, S. Ghosh, D.
S. Goodsell, R. K. Green, V. Guranovic, D. Guzenko, B. P. Hudson, T.
Kalro, Y. Liang, R. Lowe, H. Namkoong, E. Peisach, I. Periskova, A.
Prlic, C. Randle, A. Rose, P. Rose, R. Sala, M. Sekharan, C. Shao, L.
Tan, Y.-P. Tao, Y. Valasatava, M. Voigt, J. Westbrook, J. Woo, H. Yang,
J. Young, M. Zhuravleva, C. Zardecki, RCSB Protein Data Bank: Biological
macromolecular structures enabling research and education in
fundamental biology, biomedicine, biotechnology and energy. Nucleic Acids Res. 47, D464–D474 (2019).
47
E.
C. Meng, E. F. Pettersen, G. S. Couch, C. C. Huang, T. E. Ferrin, Tools
for integrated sequence-structure analysis with UCSF Chimera. BMC Bioinformatics 7, 339 (2006).
48
R. Adamczak, A. Porollo, J. Meller, Combining prediction of secondary structure and solvent accessibility in proteins. Proteins 59, 467–475 (2005).
49
A. Fiser, A. Sali, ModLoop: Automated modeling of loops in protein structures. Bioinformatics 19, 2500–2501 (2003).
50
B. Webb, A. Sali, Protein structure modeling with MODELLER. Methods Mol. Biol. 1654, 39–54 (2017).
(0)eLetters
No eLetters have been published for this article yet.
eLetters is an online forum for ongoing peer review. Submission of eLetters are open to all. eLetters are not edited, proofread, or indexed. Please read our Terms of Service before submitting your own eLetter.
Current Issue
LATEST NEWS
© 2021 American Association for the Advancement of Science. All rights reserved. AAAS is a partner of HINARI, AGORA, OARE, CHORUS, CLOCKSS, CrossRef and COUNTER. Science Advances eISSN 2375-2548.
https://www.science.org/doi/10.1126/sciadv.abf4468#.YUnkF4i9SRh.linkedin
No hay comentarios:
Publicar un comentario