2023 Article 5991
2023 Article 5991
2023 Article 5991
Open access Eukaryotic cells can undergo different forms of programmed cell death, many of
Check for updates which culminate in plasma membrane rupture as the defining terminal event1–7.
Plasma membrane rupture was long thought to be driven by osmotic pressure, but it
has recently been shown to be in many cases an active process, mediated by the protein
ninjurin-18 (NINJ1). Here we resolve the structure of NINJ1 and the mechanism by which
it ruptures membranes. Super-resolution microscopy reveals that NINJ1 clusters into
structurally diverse assemblies in the membranes of dying cells, in particular large,
filamentous assemblies with branched morphology. A cryo-electron microscopy
structure of NINJ1 filaments shows a tightly packed fence-like array of transmembrane
α-helices. Filament directionality and stability is defined by two amphipathic α-helices
that interlink adjacent filament subunits. The NINJ1 filament features a hydrophilic side
and a hydrophobic side, and molecular dynamics simulations show that it can stably
cap membrane edges. The function of the resulting supramolecular arrangement was
validated by site-directed mutagenesis. Our data thus suggest that, during lytic cell
death, the extracellular α-helices of NINJ1 insert into the plasma membrane to
polymerize NINJ1 monomers into amphipathic filaments that rupture the plasma
membrane. The membrane protein NINJ1 is therefore an interactive component of the
eukaryotic cell membrane that functions as an in-built breaking point in response to
activation of cell death.
NINJ1 is a 16-kDa plasma membrane protein that is evolutionarily con- PMR was quantified by the release of lactate dehydrogenase (LDH),
served and found in all higher eukaryotes. It has been predicted to which is too large to be released directly through GSDMD pores11–13.
feature two transmembrane helices, with the termini located in the Of note, following inflammasome activation, the amount of released
extracellular space8 (Extended Data Fig. 1a). During inflammasome- LDH increased slowly with time, whereas NINJ1 polymerization was
driven pyroptosis, NINJ1 induces plasma membrane rupture (PMR) already detectable at the onset of LDH release and increased only
downstream of the cell death executor gasdermin D (GSDMD), which marginally at later time points (Fig. 1c). Small amounts of NINJ1
in turn is activated by caspase-dependent cleavage8–10 (Extended dimers were also detectable in live cells (Fig. 1a,b), suggesting that
Data Fig. 1b–f and Supplementary Videos 1 and 2). PMR coincides higher-order NINJ1 polymers are the active species. These time-resolved
with the formation of higher-order NINJ1 polymers and membrane data are thus fully consistent with a necessity of NINJ1 polymerization
blebs8. To better understand the correlation between these events for PMR.
and to study the assembly kinetics of NINJ1 polymers in response to We also monitored NINJ1 polymerization using time-lapse fluores-
inflammasome activation, we performed crosslinking experiments cence microscopy in HeLa cells co-expressing GFP-tagged human NINJ1
in mouse bone-marrow-derived macrophages (BMDMs). In line with (hNINJ1) and a CRY2–caspase-1 fusion protein (opto-casp1), which
progressive polymerization of NINJ1, we detected the formation of enables rapid caspase activation and induction of GSDMD-driven
NINJ1 dimers and trimers 10 min after inflammasome activation; this pyroptosis in single cells using optogenetics14 (Fig. 1d and Extended
was followed by extensive polymerization of NINJ1 and the formation Data Fig. 2a–c). Quantitative inhomogeneity analysis showed that
of larger polymers at later time points (Fig. 1a,b). The bulk of NINJ1 concurring with the influx of the dye DRAQ7, which indicates a loss of
polymerization coincided with complete oligomerization of cleaved plasma membrane integrity and cell death, the diffusely localized NINJ1
GSDMD p30 (Fig. 1a,b)—that is, formation of GSDMD pores—consistent started to cluster at the plasma membrane (Fig. 1d,e, Extended Data
with NINJ1 activation occurring downstream of GSDMD activation. Fig. 2c and Supplementary Video 3), and that these clusters persisted
1
Biozentrum, University of Basel, Basel, Switzerland. 2Department of Immunobiology, University of Lausanne, Epalinges, Switzerland. 3Stuttgart Center for Simulation Science, Cluster of
Excellence EXC 2075, University of Stuttgart, Stuttgart, Germany. 4Institute of Science and Technology Austria (ISTA), Klosterneuburg, Austria. 5Department of Biosystems Science and
Engineering, Eidgenössische Technische Hochschule (ETH) Zurich, Basel, Switzerland. 6Nanoscale Infection Biology Group, Department of Cell Biology, Helmholtz Centre for Infection Research,
Braunschweig, Germany. 7Institute for Genetics, Technische Universität Braunschweig, Braunschweig, Germany. 8These authors contributed equally: Morris Degen, José Carlos Santos,
Kristyna Pluhackova. ✉e-mail: kristyna.pluhackova@simtech.uni-stuttgart.de; petr.broz@unil.ch; sebastian.hiller@unibas.ch
DRAQ7 (ΔF/Fi)
Nig: – + – +
inhomogeneity
(kDa) 30
NINJ1–GFP
40
(Dt – Di)
Ninj1–/–
Ninj1–/–
Ninj1–/–
Ninj1–/–
n-mers 20
GSDMD
WT
WT
WT
WT
250 20
10
(kDa) 55 FL
0 0
n-mers 35 p30
–10 0 10 20 30 40 50 60
250
GSDMD
Time (min)
55 FL
250 f
35 n-mers
p30
130 40 60
DRAQ7 (ΔF/Fi)
100
inhomogeneity
30
HATMD–GFP
70
NINJ1
40
(Dt – Di)
250 55 4-mer
20
n-mers 3-mer
130 35 20
2-mer 10
100 25
70 1-mer 0 0
4-mer 15
55 –10 0 10 20 30 40 50 60
NINJ1
3-mer 1-mer
15 (short exp.) Time (min)
35 2-mer
25
g
15 1-mer
n-mers
Tubulin
250 40 60
DRAQ7 (ΔF/Fi)
E-cadherin–GFP
1-mer
inhomogeneity
15 (short exp.) 130 30
40
(Dt – Di)
55 1-mer 20
20
n-mers c 80 10
LDH release
250 60
0 0
Tubulin
(%)
40
–10 0 10 20 30 40 50 60
20
0 Time (min)
55 1-mer
Nig (min): 0 7 10 20 30 60 90
d Time (min)
–7 0 2 5 8 16 39 54
Fig. 1 | Polymerization kinetics of plasma membrane NINJ1. a,b, Western to track plasma membrane permeabilization. Time was normalized to the onset
blot analysis of endogenous GSDMD and NINJ1 in primed BMDMs after nigericin of increase in DRAQ7 nuclear fluorescence. White arrows indicate regions that
stimulation (Nig) for 1.5 h (a) or for 2, 5, 15, 25, 55 or 85 min (b) followed by are enlarged in the insets. Scale bar, 10 µm. e–g, Normalized quantification of the
treatment with the membrane-impermeable BS3 crosslinker for 5 min. FL, full distribution inhomogeneity of NINJ1–GFP (e), HATMD–GFP (f) or E-cadherin–GFP
length. Of note, tubulin, used here as a loading control, is crosslinked owing to (g) at the basal plane of cells, and DRAQ7 nuclear fluorescence intensity after
BS3 entry through GSDMD pores under nigericin-treated conditions. Time in photoactivation of opto-casp1 (F). The distribution inhomogeneity at each
b is the total incubation time for nigericin and BS3 treatment. Short exp., short time point (D t) was normalized to the distribution inhomogeneity at the initial
exposure. For gel source data, see Supplementary Fig. 1. c, LDH release from time point of the experiment (Di). Data are mean ± s.d. Data are representative
primed BMDMs after nigericin stimulation. d, Time-lapse fluorescence confocal of 2 (b), 3 (a) or 14 (d) independent experiments, or pooled from 2 independent
microscopy of HeLa cells co-expressing hNINJ1–GFP and opto-casp1 following experiments performed in triplicate (c) or at least 10 (e–g) independent
photo-activation. Images show the NINJ1–GFP fluorescence at the basal plane experiments.
of the cell and the influx of DRAQ7 (maximum (max.) projection from a z-stack)
beyond cell lysis. Notably, the formation of clusters during pyroptosis PMR is a cell-intrinsic process or whether this activity is dependent
was specific for NINJ1, as other plasma membrane proteins such as on neighbouring cells via released NINJ1 or direct contact. Co-culture
the haemagglutinin transmembrane domain15 (HATMD) or E-cadherin experiments of wild-type BMDMs with Casp11-deficient BMDMs trans-
did not cluster (Fig. 1f,g, Extended Data Fig. 2d,e and Supplementary fected with lipopolysaccharide (LPS) to activate the non-canonical
Videos 4 and 5). Similar assemblies were also formed by endogenous inflammasome unambiguously demonstrated that NINJ1 lyses cells
NINJ1 in wild-type but not in Gsdmd—/— BMDMs upon activation of the in a cell-intrinsic manner without affecting immediate neighbours
NLRP3 or AIM2 inflammasomes (Extended Data Fig. 3a). Clustering of (Extended Data Fig. 3c). NINJ1-driven PMR in inflammasome-activated
NINJ1 and NINJ1-dependent LDH release was also detected upon induc- cells is thus a cell-intrinsic process, which involves the formation of
tion of apoptotic cell death8 (Extended Data Fig. 3a,b), but in this case GSDMD pores and rapid NINJ1 polymerization, followed by membrane
independently of GSDMD. Next, we investigated whether NINJ1-driven rupture with slower kinetics.
Untreated
Untreated
Rg (nm)
1 2 102
2 5 101
3
1
4
10 **
+ B/B
+ B/B
3 6
Ecc
5 6 5
c e
1
Large
clusters
Medium- 2 2
sized
clusters
Small
clusters
Fig. 2 | Super-resolution imaging of NINJ1 assemblies. a, Wide-field imaging clusters found in pyroptotic cells. The small clusters are also observed in
of DmrB–Casp4tg HeLa cells expressing hNINJ1–GFP and GSDMD used for non-activated cells. Scale bars, 500 nm. d, Radius of gyration (Rg) and eccentricity
STORM microscopy in b–e. Cells were left untreated or stimulated with B/B (Ecc) for each identified hNINJ1 cluster. Plots show the distribution of all identified
homodimerizer 3 h before fixation and labelling with Alexa Fluor 647-conjugated clusters from three independent experiments. The lines indicate median values.
anti-GFP nanobodies. Scale bar, 50 µm. b–e, STORM super-resolution imaging of Statistical analysis based on the median Rg and Ecc of each experiment using
hNINJ1–GFP in cells from a using TIRF illumination of the basal plane. b, Left, Student’s unpaired two-sided t-test. **P < 0.01, ***P < 0.001. e, Overview STORM
untreated or B/B-stimulated cells expressing hNINJ1–GFP. Scale bar, 10 µm. reconstruction of assemblies in B/B-stimulated cells expressing hNINJ1–GFP
b, Right, STORM super-resolution reconstruction of hNINJ1–GFP labelled with including filamentous structures. Two filaments are highlighted with magenta
Alexa Fluor 647-conjugated anti-GFP nanobodies. The indicated outlined arrowheads. The indicated regions are magnified on the right. Scale bars, 1 µm.
regions are magnified on the right. Scale bar, 500 nm. c, Gallery of hNINJ1–GFP Data in a–c,e are representative of at least three independent experiments.
60 Å
2D classes
c K65
C terminus I α3 α4 G124 II α2 α3 III
α2
N60 V68
F135
L56 V67 L80
V141 M51
Kink L56
L64 I84 V131 L88
K44
α1 N60 I92
α4 α1 L88 F127
V96
α3 L56 Q91
D53
N120 I99
S43 A47 a1 V94
N39 K45
N terminus I54 I99 α4 α3
d e
50 ****
****
Permeability (%)
180° 40
30
20
10
0
1: 0
1: 0
0
y
Hydrophilic Hydrophobic
48
24
12
pt
Em
1:
NINJ1:lipid ratio
180°
Acidic Basic
Fig. 3 | Cryo-EM structure of NINJ1 filament. a, Cryo-EM micrograph showing comprising helices α1–α4, with surface representation outlined in light grey.
filamentous hNINJ1 (white arrows) along representative 2D classes. Scale bar, d, Lipophilicity and charge distribution of the hNINJ1 filament. e, Permeability
25 nm. Micrograph representative of 13,124 micrographs from one dataset. of hNINJ1 proteoliposomes at different protein:lipid molar ratios. Data are
b, Organization of hNINJ1 filaments with helices represented as tubes and each mean + s.d. (n = 3 independent experiments). Statistical analysis by one-way
subunit shown in a colour gradient (yellow–green–purple). The main interaction ANOVA. ****P < 0.0001.
interfaces I, II and III are shown below. c, A single hNINJ1 filament subunit,
Fig. 5c–h). The filaments were linear stacks of subunits, with an interval conserved residues K45 and D53 is the most prominent (Fig. 3b, I).
of 20.95 Å and a slight rotation of –1.05° per subunit. The cryo-EM map The interaction also includes newly formed intramolecular contacts
had a resolution of 3.8 Å, enabling us to build a molecular model of via an extensive hydrophobic patch on the amphipathic helix α2 that
filamentous hNINJ1 (Fig. 3a,b and Extended Data Table 1). The first 38 matches a complementary side chain array on α3 (Fig. 3b, II). Finally,
residues of hNINJ1 remained disordered, as shown by solution NMR hydrophobic residues on α3 align with a complementary set of residues
relaxation experiments (Extended Data Fig. 6a), in full agreement with on α4 of the neighbouring protomer, presumably including a cation–π
predictions from AlphaFold and previous NMR experiments19,20. Indeed, interaction between K65 and F135 (Fig. 3b, III).
a truncation experiment showed these residues to be dispensable for Notably, the experimental structure of NINJ1 in the filament over-
hNINJ1 filament formation (Extended Data Fig. 6b). The next 103 resi- laps nearly perfectly with the AlphaFold model for helices α2, α3
dues (residues 39–141) were well represented in the maps and could and α4, but not for helix α1 (Extended Data Fig. 6c). In the AlphaFold
be modelled unambiguously as four α-helices α1–α4 (Extended Data model, α1 and α2 combine to form a single, straight helix. We used
Fig. 5g). Notably, the experimental density comprised two identical molecular dynamics simulations of hypothetical filaments, in which
hNINJ1 filaments, which were packed together in an antiparallel arrange- we replaced the individual monomers in the experimentally deter-
ment (Extended Data Fig. 5h). mined structure with the AlphaFold model. In several subunits, the
The hNINJ1 filament is organized by stacking of adjacent protomers single α1–α2 helix began to kink at residue 56 and helix α1 restruc-
in a fence-like manner (Fig. 3b). The two antiparallel helices α3 and tured itself towards the cryo-EM structure (Extended Data Fig. 6d).
α4 (residues 79–103 and 114–138, respectively) form the core of the Furthermore, we analysed co-evolution of NINJ1, which also underlies
filament. These helices are hydrophobic and form a hairpin of trans- the model building in AlphaFold, and focused on the 100 most sig-
membrane helices in the inactive form of the protein (Extended Data nificant co-evolution pairs. A large majority of these residues pairs
Fig. 1a). The two N-terminal helices α1 and α2 (residues 44–55 and 58–74, corresponds to intramolecular short-range contacts within the NINJ1
respectively) are separated by a distinct kink at L56 (Fig. 3c). α2 thus monomer. Nine of the residue pairs, however, are in closer spatial
adopts a parallel orientation with respect to α3 and α4, whereas α1 proximity between subunits in the filament structure than within
protrudes at a nearly 90° angle from the helical bundle and connects one monomer (Extended Data Fig. 6e). In particular, this includes the
to the adjacent protomer via an extensive polymerization interface. residue pair F127–G95, which has the highest significance score of all
The intermolecular contacts include multiple side chain interactions pairs in NINJ1 and which intermolecularly connects helices α3 and α4.
between helix α1 of one protomer and helices α1, α3 and α4 of the These evolutionary data lend further support for the relevance of
neighbouring protomer, of which a salt bridge between the highly filamentous NINJ1 in a biological context.
A138
T123
L121
I134
Q91
G95
D53
K44
K45
A47
V82
I84
Q91 c In vitro
K44 G95 filaments: + + – + – + + – – – + + – – –
L121
K45 I134 NS ***
d e
L121W
Mock
WT
A138L
***
L121F
T123L
V82W
K44Q
K45Q
Q91A
I134F
D53A
G95L
A47L
V82F
I84F
50 **** 80 WT Ninj1–/–
Permeability (%)
40
LDH release)
Cell lysis (%
60 (HEK 293T + mNINJ1)
30 NS NS NS f 250 NS
40
****** **
WT
WT BMDMs
–
Vector
A47L
I84F
G95L
K45Q
D53A
Q91A
V82F
V82W
50
0
Ninj1–/– + mNINJ1
Mock
WT
L121W
A138L
L121F
T123L
V82W
K44Q
K45Q
Q91A
I134F
D53A
G95L
A47L
V82F
I84F
(BMDMs + nigericin)
Cytosol
NINJ1 activation
Formation
of large
polymers
Fig. 4 | The mechanism of NINJ1-mediated PMR. a, Three subunits of to the control condition highlighted in bold. Data are mean + s.d. and
filamentous hNINJ1 with overview of residues selected for mutagenesis study data are pooled from two independent experiments performed in triplicate
(intermolecular interactions, magenta; intramolecular interaction, purple; (c; for K45Q, K45Q, A47L, D53A, V82W, L121W, T123L and A138L mutants),
membrane interactions, green). b, Schematic representation of the residues three independent experiments performed in triplicate (c; for mock, WT
selected for mutagenesis. c, Cytotoxicity upon overexpression of wild-type and V82F, I84F, Q91A, G95L and A138L mutants), and representative of two
(WT) or mutant mNINJ1 in HEK 293T cells. d, Permeability of proteoliposomes independent experiments performed in triplicate (c; for L121F mutant), pooled
containing wild-type and mutant mNINJ1 compared with protein-free from two independent experiments performed in triplicate (f) or representative
liposomes (empty). Data are mean + s.d. (n = 3). e, Release of LDH in primary of two independent experiments performed in triplicate (e). In c,e,f, multiple
Ninj1–/– BMDMs reconstituted with wild-type mNINJ1 or different mNINJ1 plates were used to test all mutants, thus control conditions were included in
mutants upon nigericin treatment (1.5 h). Reconstitution with the empty each of the plates. *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001; NS, not
vector and non-transduced Ninj1–/– BMDMs (–) were used as controls. significant. P values by one-way ANOVA with Dunnet’s multiple comparisons
f, Cytotoxicity upon B/B treatment in HeLa cells co-expressing DmrB–caspase-4 tests. g, Structural model of NINJ1-mediated membrane rupture. Non-activated
and wild-type or mutant mNINJ1. Killing score corresponds to the cytotoxicity, NINJ1 is randomly distributed in the plasma membrane (PM). Upon activation,
measured by LDH release, normalized against wild-type mNINJ1 control (c) or NINJ1 polymers lyse the membrane, resulting in the release of cytosolic
mock-treated controls (f). Statistical analysis in c–f by individual comparison content (red).
molecular patterns (DAMPs) and other cellular content are released into the GSDMD pore has a limited pore size that allows interleukin release
the extracellular milieu. Although single filaments might be sufficient while retaining larger molecules13, the membrane openings or lesions
to damage membranes, it is also plausible that NINJ1 forms double caused by NINJ1 filaments are essentially unconstrained in size. NINJ1
filaments that open up in a zipper-like manner in response to osmotic lesions appear functionally related to the large superstructures of acti-
pressure to form membrane lesions. vated mitochondrial Bax and Bak, in that they both lyse membranes30.
The trigger that causes the transition of NINJ1 from the inactive Although atomic structures of Bax or Bak in a pore-forming conforma-
state to the active state remains unknown. The proposed polymeriza- tion are not currently available, we speculate that the helical hairpin of
tion mechanism raises the interesting possibility that the membrane α3 and α4 in the NINJ1 filament might show functional and structural
composition might contribute at least partially an activation signal. resemblance to the helical hairpin of α5 and α6 in activated Bax. NINJ1
During cell death, negatively charged phosphatidylserine becomes was initially reported as an adhesion molecule, induced after sciatic
exposed on the cell surface28,29, which might be recognized by helices α1 nerve injury and promoting axonal growth31–33. Given that NINJ1 drives
and α2 of NINJ1. Indeed, lipid binding experiments and dye release cell death and the release of DAMPs, and close links exist between cell
assays show that a peptide corresponding to helices α1 and α2 inter- death, inflammation and tissue repair34, it is conceivable that NINJ1 has
acts specifically with POPS-containing membranes, and molecular an indirect role in promoting axonal growth either by causing inflam-
dynamics simulations show the same effect (Extended Data Fig. 9a–c). mation or the release of stimulatory molecules. Conversely, it is also
Membrane-composition sensing as a potential activation mechanism possible that NINJ1 has a dual role, serving in both cell–cell adhesion
of NINJ1 is thus a promising avenue for future work. and as a breaking point for membranes at strong osmotic pressure35.
In summary, active NINJ1 has a unique structure with long, α-helical The structure, along with the mutagenesis studies provide possible
filaments capping membrane edges. Whereas the β-sheet structure of explanations for why NINJ2 is not able to functionally replace NINJ1,
Extended Data Fig. 3 | Cell death triggers NINJ1 polymerization at CFSE. Co-cultured cells were then transfected with purified LPS and the
the plasma membrane and induces PMR in a cell-intrinsic manner. incorporation of propidium iodide (PI) in dying cells (dashed contour)
a, Immunofluorescence microscopy of endogenous NINJ1 in primed BMDMs was quantified in both cell types by fluorescence microscopy. Different
upon stimulation with Nigericin (1 h), or poly(dA:dT) transfection (1 h) (to co-cultures were prepared (WT + WTCFSE-pos; C11–/– + C11–/–CFSE-pos; WT + C11–/–CFSE-pos;
induce pyroptosis), or in naïve cells stimulated with TNF + SM (SMAC-mimetic C11–/– + WTCFSE-pos), and the percentage of cells with PI positive nuclei was quantified
AZD 5582) for 16 h (to induce apoptosis). Scale bars, 10 µm. b, LDH release in in the CFSE-negative and -positive cells. Graphs show the mean ± SD and data
WT, Gsdmd–/– or Ninj1–/– BMDMs stimulated with TNF + SM (16 h). c, Schematic are representative from two independent experiments (a), or pooled from
representation of the experimental setup to assess if NINJ1-driven PMR acts two independent experiments performed in triplicate (b,c). *** P < 0.001; ns,
in a cell-intrinsic (in Cis) or -extrinsic (in Trans) manner: primed WT BMDMs not significant. P-values were calculated using one-way ANOVA with multiple
were co-cultured with Casp11–/– (C11–/–) BMDMs previously incubated with comparisons Dunnett tests.
Extended Data Fig. 4 | See next page for caption.
Article
Extended Data Fig. 4 | NINJ1-GFP cluster size and resolution after density- interest (ROIs) across multiple cells, with equal total area for each condition.
based clustering. a, Release of LDH in HeLa cells expressing DmrB-Casp4 upon Rg and FRC are plotted against the number of localizations per cluster for each
activation with different doses of B/B homodimerizer for 3 h. Graph shows the identified cluster. For plotting and further analysis in panels d–f, only clusters
mean ± SD and data are representative from two independent experiments with more than 100 localizations were used. d, Histogram of cluster sizes (Rg).
performed in triplicate. b, Absence of reorganization of the control protein e, Distribution of cluster per area cell surface. Analysed were 23/72 cells from
HATMD-GFP upon cell death. HeLa cells co-expressing DmrB-Casp4 and 2/4 independent experiments (UT/+BB). Lines indicate mean values. f, Extended
HATMD-GFP 15 were untreated (UT) or stimulated with B/B homodimerizer and gallery of large NINJ1-GFP structures found in pyroptotic cells. Gallery of clusters
imaged using TIRF microscopy to visualize the membrane organization of identified after stimulation (as in Fig. 2c) and selected by the cyan quadrant
HATMD-GFP. c, Density-based clustering of NINJ1 clusters, correlating their FRC (Rg > 200 nm, localizations > 500) in c. Scale bar, 2 µm.
resolution and size (Rg). Clustering was performed in multiple regions of
Extended Data Fig. 5 | Structure determination of hNINJ1 filaments. d, Cryo-EM processing workflow. e, Distribution of orientations over azimuth
a, Negative Stain TEM micrograph of purified hNINJ1 ring assemblies (white and elevation angles for particles included in the calculation of the final map
arrows). Scale bar, 50 nm. Micrograph representative of n = 14 independent (n = 709,840 particles). f, Gold standard Fourier shell correlation (GSFSC)
experiments. b, Mass photometry measurements of purified hNINJ1 in presence plot for final helical refinement of hNINJ1 dataset. g, Modeled helices α1−α4
of detergent. c, Western Blot of purified hNINJ1 sample used for cryo-EM shown individually within the experimentally determined cryo-EM density
structure determination. For gel source data, see Supplementary Fig. 3. (grey mesh). h, Collection of views of the hNINJ1 double filament model.
Article
Extended Data Fig. 6 | Structural and functional analysis of NINJ1. a, 15N R 2 α3 and α4 was restrained to study the propagation of helices α1 and α2. e, Residue
transverse relaxation rate measurements of GB1-hNINJ1–82 in the absence of any pairs with co-evolutionary coupling displayed on the filament structure of hNINJ1.
membrane-mimicking environment show that the N-terminal ~40 residues are Shown are all pairs among the 100 most significant couplings, that are closer in
highly flexible, while the remainder has limited flexibility. b, Negative stain space in the filament than in the monomer. A list of these pairs is given in
micrograph of purified hNINJ1∆1–36 showing ring assemblies (white arrows), Supplementary Table 2. f, Hydrophobic interaction occurring between hNINJ1
scale bar corresponds to 50 nm. Micrograph representative of n = 2 independent filaments. g, Relative fluorescence traces of hNINJ1 proteoliposomes with
experiments. c, Overlay of 20 Alphafold predictions of hNINJ1 with the different lipid to protein molar ratios, as indicated. Star indicates the addition
experimentally determined structure. d, Snapshots at 0 µs and 1 µs of an AA of dithionite. Arrow indicates timepoint of Triton X-100 addition. Triplicates
simulation of a hypothetical filament, where the experimental monomer was are shown.
replaced by the Alphafold-predicted monomer (n = 1). The backbone of helices
Extended Data Fig. 7 | See next page for caption.
Article
Extended Data Fig. 7 | Stability of NINJ1 polymers probed by molecular snapshot after 40 µs of a CG simulation of a circular hNINJ139–152 45-mer in a lipid
dynamics simulation. a, Snapshot after 1 µs of an all-atom simulation of linear bilayer (n = 4). Insets on the right show the protein packing in a straight (bottom)
NINJ1 oligomers on both edges of a membrane stripe (n = 2). b, Snapshot after and in a curved part of the polymer (top). g, Time evolution of the average
20 µs of a CG MD simulation of linear filaments capping the edges of a membrane inner area of rings made of 45 hNINJ139–152 (n = 4) or hNINJ176–152 (n = 3) relative
stripe (n = 2). c, Root mean square deviations (RMSD) over simulation time of to the area at t = 3 µs. The shaded areas show SEM over individual simulations.
the backbone of linear pentamers attached to the membrane patches. Average h, Snapshots at 150 µs of four independent CG MD simulations of 45-mer
RMSD and SEM over individual, unique pentamers and simulations (n = 4 for AA hNINJ139-152 (n = 4). i, Starting structure, collapse at 3 µs, as well as three snapshots
and n = 8 for CG) are given. d, Average structure and B-factors (in Å 2) of the at 150 µs of independent CG MD simulations of 45-mer hNINJ176–152 rings, i.e.
hNINJ1 39–152 backbone (n = 20) in simulations (a). Structures were aligned to missing helices α1 and α2 (n = 3). Lipids are shown as grey sticks, with
helices α3 and α4 prior analysis. e, Representative snapshots of CG MD phosphates highlighted as spheres. hNINJ1 protomers are randomly colored
simulations of the hNINJ1 39–152 double filament after 50 µs in absence (left, n = 2) yellow, green or purple with the exception of c, where the purple-green-yellow
and presence (right, n = 2) of DDM. The black box highlights the simulation unit coloring scales with the B-factor (purple low B-factor, yellow large B-factor).
cell. Periodic filament copies were added on right and left. f, Representative
Extended Data Fig. 8 | See next page for caption.
Article
Extended Data Fig. 8 | Analysis of NINJ1 single-point mutants. a, Comparison Dox-inducible mNINJ1 and induction of protein expression for the indicated time
of hNINJ1 and mNINJ1 sequences. Orange background shows identical residues. points. g, LDH release in HeLa cells co-expressing Dox-inducible DmrB-Casp4
Secondary structure is indicated on top. In the structured part (residues 44–138), and Dox-inducible WT or different mNINJ1 mutants. Cells were treated with Dox
2 amino acids differ and 93 are identical. Arrows indicate selected single residues for 16 h, followed by treatment with B/B homodimerizer. B/B untreated cells were
that were mutated in this work. b, c, Western blot analysis of mNINJ1 also used as control. h, Western blot analysis of mNINJ1 and GSDMD expression
expression (b) and LDH release (c) in HEK 293T cells transiently transfected in HeLa cells stably co-expressing FLAG-GSDMD-V5 and Dox-inducible DmrB-
with WT or different mNINJ1 mutants. mNINJ1 expression was induced by Casp4, and transiently expressing WT or different mNINJ1 mutants. i, Western
adding doxycycline (Dox) for 48 h. d, Triplicates of fluorescence traces of WT- blot analysis of GSDMD processing in HeLa cells stably co-expressing FLAG-
NINJ1, mutants, and empty liposomes. Star indicates the addition of dithionite. GSDMD-V5 and Dox-inducible DmrB-Casp4, and transiently expressing WT or
Arrow indicates timepoint of Triton X-100 addition. e, Western blot analysis of different mNINJ1 mutants, upon caspase-4 activation with B/B homodimerizer.
mNINJ1 expression in Ninj1–/– primary BMDMs transduced with a retroviral vector Cell lysates and supernatants were combined and analyzed. In (h,i) mNINJ1
expressing WT or different mNINJ1 mutants. Transduction with the empty vector expression was induced with Dox for 16 h. Graphs show the mean ± SD and data
or non-transduced Ninj1–/– cells (–) were used as controls. f, LDH release and are representative from at least two independent experiments performed in
western blot analysis of mNINJ1 expression in HeLa stably expressing FLAG- triplicate. For gel source data, see Supplementary Fig. 4.
GSDMD-V5 and Dox-inducible DmrB-Casp4, upon transient transfection with
Extended Data Fig. 9 | Interactions of the NINJ1 N-terminal helices * P < 0.05, ** P < 0.01, *** P < 0.001; ns, not significant. c, Residue-specific
with membranes. a, Quartz crystal microbalance with dissipation membrane interaction probability of hNINJ1 20–152 from CG MD simulations
(QCM-D) experiments. Frequency change upon addition of GB1-hNINJ1(1–81) (n = 10). Membrane with PS only in the cytosolic leaflet, corresponding to a
to supported lipid bilayer membranes containing only DOPC (black), healthy cell, are shown in green, while membrane with PS in both leaflets,
80% DOPC/20% DOPS (orange) or 60% DOPC/40% DOPS (red). Duplicate corresponding to activated cell death signal, is shown as grey bars.
experiments with independently produced protein and lipid bilayers are The N-terminal adhesion motif (NAM) is highlighted. The error bars denote
denoted with squares and circles. The lines denote fits of a Hill equation. SEM over the individual simulations (n = 10). Significance was calculated
b, Liposome composition-dependent dye release by hNINJ140–60 peptide or its using Student’s paired two-sided t-test. * P < 0.05, ** P < 0.01; *** P < 0.001.
scrambled sequence version. Data are means of n = 4 replicates for 3.75 µM d, Sequence alignment of hNINJ1 and hNINJ2. Conserved residues are
and n = 5 replicates for the other concentrations, with error bars denoting SD. highlighted in orange. The secondary structure of hNINJ1 is shown on top.
Significance was calculated using Student’s unpaired two-sided t-test.
Article
Extended Data Table 1 | Cryo-EM data collection, refinement and validation statistics
nature portfolio | reporting summary
Sebastian Hiller, Petr Broz, Kristyna
Corresponding author(s): Pluhackova
Last updated by author(s): Mar 7, 2023
Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.
For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.
For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.
Data analysis NMR: CCPN version 3.0.4, xmgrace version Grace 5.1.25; MD: GROMACS 2020/2021, R version 3.6.3, PyMol 2.5.0; STORM: Data analysis:
Decode v0.1, Matlab 2020b, Fiji with ImageJ 1.53q, ThunderSTORM v1.3, own custom processing codes (https://github.com/christian-7/); Cell
biology: Microsoft Excel for Mac v16, Prism Graphpad 9.0, Gen5, FiJi, Zen Blue 2.3 Imaging Software; Structural Biology: Phenix Version
1.19.2-4158; Coot 0.8.9.2 EL, 0.9.7 EL; cryoSPARC v3.2, v3.3.2; ChimeraX v1.2.5,1.3, Alphafold 2.2.0, EvCouplings V2 webserver, MolProbity
v4.5.2 Webserver
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
March 2021
1
nature portfolio | reporting summary
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy
The atomic coordinates of filamentous hNINJ1 have been deposited in the RCSB Protein Data Bank with the accession code 8CQR. The cryo-EM map has been
deposited in the EMDB with accession code EDB-16799. Source data are provided with this paper. All other data that support the findings of this study are available
from the corresponding authors upon request. Explanation: This combination of Data Bank Deposition / Source Data Provided / Data upon request is the most
efficient combination of data forsharing in terms of curation recources spent vs. user access frequency. Protein sequences were retrieved from the Uniprot
database: Q92982 - hNINJ1, O70131 - mNINJ1, Q9NZG7 - hNINJ2.
Population characteristics NA
Recruitment NA
Ethics oversight NA
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf
Replication Each experiment was repeated as described in the figure legends. Each replicable experiment was repeated at least three times
independently. The number of replicate performed for each figure is stated in the figure legends. Experiments were repeated, when possible,
by different experimenters to ensure the reproducibility.
Randomization There was no randomization for these experiments. This study is not a randomised control trial and randomisation is not conventionally used
in in vitro/in cellulo studies such as this one. All groups of experiments were performed using the same experimental conditions and
protocols.
March 2021
Blinding Assays used objective quantification methods that are not susceptible to bias, so samples were not blinded.
2
Materials & experimental systems Methods
Antibodies
Antibodies used Antibodies used for western blot:
Mouse monoclonal anti caspase-4 clone 4B9 (Enzo life Sciences, cat #ADI-AAH-114-5; dilution 1:1000)
Mouse monoclonal anti-V5 (ThermoFischer Scientific, cat# R960-25; dilution 1:2000)
Rabbit monoclonal anti-mouse NINJ1 (rabbit IgG2b clone 25, a kind gift from Genentech, no catalogue number; dilution 1:8000)
Rabbit monoclonal anti-GSDMD (EPR19828, Abcam, cat# ab209845; dilution 1:1000)
Mouse monoclonal anti-Tubulin clone DM1A (Abcam, cat# ab40742; dilution 1:2000)
Goat anti-Rabbit IgG-HRP (SouthernBiotech, cat# 4030-05; dilution 1:5000)
Goat anti-Mouse IgG(H+L), Rat ads-HRP (SouthernBiotech, cat#1034-05; dilution 1:5000)
Mouse anti-human NINJ1 (BD Transduction Laboratories cat#610777, lot#1070002; dilution 1:1000)
Goat anti-mouse IgG HRP conjugate (MilliporeSigma cat#12-349, lot#3722026; dilution 1:2000)
Validation Anti-Caspase-4, anti-GSDMD and anti-NINJ1 antibodies were validated using knockout cell lines, validated by the suppliers and are
extensively used in the scientific community.
Anti-V5 was validated by overexpressing protein tagged V5-epitope and have been validated by their respective manufacturers. Anti-
tubulin has been validated by their manufacturer and is extensively used in the scientific community.
Authentication Cell lines were obtained from ATCC and authenticated by the vendor. The identity of cell lines was frequently checked by
their morphological features and did not show any signs of cross-contamination.
Mycoplasma contamination Cell lines are regularly tested in the lab for mycoplasma contamination and are mycoplasma free.
Commonly misidentified lines No commonly misidentified cell lines were used in this study.
(See ICLAC register)
Laboratory animals C57BL/6 mice were used in this study to generate bone marrow-derived macrophages. The specific strains (WT, Ninj1-KO and
Gsdmd-KO) are indicated, and BMDMs were harvested from both male and female 8-10 week old mice. All mice were bred and
housed at a specific-pathogen-free facility at 22 +/- 1 C° room temperature, 55 +/- 10 % humidity and a day/night cycle of 12h/12h at
the University of Lausanne.
March 2021
Reporting on sex The findings apply to both females and males. No differences have been observed between BMDMs from male or female mice
Field-collected samples Study did not involve samples collected from the field
Ethics oversight All experiments involving animals were performed under the guidelines and approval from the cantonal veterinary office of the
Canton of Vaud (Switzerland). License number VD3257
3
Note that full information on the approval of the study protocol must also be provided in the manuscript.