Paper - Review

DOI: 10.1038/s41467-020-20320-y

Abstract

1⃣ NLRP1 2⃣ CARD8
→ are related ← cytosolic sensors
← which upon activation form supra-molecular signaling complexes → known as "canonical inflammasomes"
→ resulting → in 1⃣ caspase-1 activation 2⃣ cytokine maturation 3⃣ pyroptotic cell death

1⃣ NLRP1 2⃣ CARD8
→ use their C-terminal (CT) fragments
← which containing 1⃣ a caspase recruitment domain (CARD) sub-domain 2⃣ the UPA (conserved ← in 1⃣ UNC5 2⃣ PIDD 3⃣ ankyrins) subdomain → for self-oligomerization
→ to recruit 1⃣ the inflammasome adaptor ASC 2⃣ caspase-1

Report → cyro-EM structures ← of 1⃣ NLRP1-CT 2⃣ CARD8-CT assemblies
← which the respective CARDs from central (helical filaments) ← that are promoted by oligomerized

The UPA → reduces
→ the threshold ← needed for 1⃣ NLRP1-CT 2⃣ CARD8-CT filament formation & signaling

1⃣ The mode of ASC recruitment ← by NLRP1-CT 2⃣ The contrasting direct recruitment ← of caspase-1
∵ Structural analyses

Subunits
← in the central NLRP1-CARD filament
→ dimerize ← with additional exterior CARDs
← which 1⃣ doubles its thickness 2⃣ is unique ← among all known CARD filaments

Determine
→ the structure ← of ASC-CARD-caspase-1-CARD octamer
∴ ASC → uses opposing surfaces → for NLRP1

Introduction

Innate immune pathways
→ (recognize & respond) → to a diverse array of (intra-cellular threats)

Cells
→ (identify & amplify) → danger signals
→ through supra-molecular signaling complexes
∴ Canonical inflammasomes

Sensor proteins
→ facilitate → inflammasome assembly
← by undergoing large conformational changes
← that lead to their oligomerization

Nucleotide binding domains (NBDs)
→ often drive → this self-oligomerization
→ to cluster death-fold domains
← recruit 1⃣ downstream adapter 2⃣ effector molecules

These death-fold domains
← e.g. 1⃣ pyrin domains (PYD) 2⃣ caspase recruitment domains (CARD)
→ participate ← in homotypic (1⃣ CARD-CARD 2⃣ PYD-PYD) interactions
→ lead to polymerization of caspase-1 into filaments

This process
→ increases → the local concentration of (caspase catalytic domains)
→ to facilitate 1⃣ its homo-dimerization 2⃣ auto-proteolysis
∴ Resulting → in its activation

Active caspase-1
→ processes → pro-inflammatory cytokines ← e.g. 1⃣ pro-IL-1β 2⃣ pro-IL-18 → to their bioactive forms
→ cleaves → the pore-forming protein gasdermin-D (GSDMD) → to 1⃣ promote cytokine release 2⃣ concomitant (lytic cell death) → pyroptosis

1⃣ NLRP1 2⃣ CARD8
← two related inflammasome sensors
→ are highly expressed ← in a number of (cell types)
→ play important roles ← in both 1⃣ host defense 2⃣ human diseases

Keratinocytes
→ express → high level of 1⃣ NLRP1 2⃣ germline mutations of diseases
→ to a number of skin-related inflammatory diseases

These mutations
→ cause → constitutive 1⃣ NLRP1 activation 2⃣ downstream pyroptosis
→ leading to (damaging inflammation)

❗: Understanding (the molecular mechanisms)
← governing 1⃣ NLRP1 2⃣ CARD8 inflammasome signaling
→ will facilitate → the discovery of new therapeutics
→ for inflammatory diseases & cancers

1⃣ NLRP1 2⃣ CARD8
→ mediate → inflammasome formation
→ through their CARD-containing C-terminal fragment (CT)
→ generated upon (functional degradation) of (their respective N-terminal fragment)

This unusual mechanisms of (inflammasome activation)
→ is associated ← with auto-proteolysis of the function-to-find domain (FIIND)
→ results ← in non-covalently associated (NT & CT)

1⃣ NLRP1-CT 2⃣ CARD8-CT
→ are repressed ← by the NT ← until upstream cues induce proteasomal degradation of the NT
→ concomitant release of the CT

One such cue
→ is provided ← by small molecule inhibitors
← of di-peptidyl peptidases 1⃣ DPP9 2⃣ DPP8

❓: the mechanisms ← of 1⃣ NLRP1 2⃣ CARD8 activation
→ is currently unclear

Activated 1⃣ CARD8 2⃣ NLRP1 inflammasomes
→ are distinctive
← in they are only composed of 1⃣ UPA 2⃣ CARD

Other inflammasomes
← e.g. 1⃣ NLRC4 2⃣ NLRP3
→ use 1⃣ NBDs 2⃣ leucine-rich repeats (LRRs)
→ to facilitate 1⃣ oligomerization 2⃣ inflammasome activation

The specificity of 1⃣ NLRP1-CT 2⃣ CARD8-CT
→ for the 1⃣ CARD 2⃣ PYD containing adaptor ASC
→ for the caspase-1
→ differ ← from most other (inflammasome sensors)

ASC
→ is a nearly universal adapter
← that bridges interaction ← between 1⃣ sensors 2⃣ the CARD-containing caspase
CARD8
→ does NOT interact ← with ASC
→ directly engages caspase-1

NLRC4
← the CARD-containing sensor protein
→ can activate caspase-1 → both (w/ & w/o ASC)
Human NLRP1
→ is completely dependent ← on ASC
→ cannot ❌ engage → caspase-1 directly

❓: The molecular bases
→ for the assembly of 1⃣ UPA-CARD-mediated NLRP1-CT 2⃣ CARD8-CT inflammasomes
→ for their differential specificity → to 1⃣ ASC 2⃣ caspase-1
→ are unknown

❗: determined → the cyro-electron microscopy (cryo-EM) structure ← of 1⃣ NLRP1-CT 2⃣ CARD8-CT filaments
❗: analyzed → their assembly & specificity

The UPA domain → is required
→ for productive 1⃣ NLRP1 2⃣ CARD8 inflammasome signaling
∴ Promote → 1⃣ CARD clustering 2⃣ filament formation
∴ Serve → an analogous function → to the (NBD & LRRs) ← in many other inflammasome sensors

The CARD filament of NLRP1-CT
→ is composed ← of CARD dimers
← which an additional CARD flanks → each sub-unit of the central CARD filament

Determined → the structure of an ASC-CARD-caspase-1-CARD octamer
∴ ASC uses → opposing surfaces → for caspase-1 🆚 NLRP1 engagement
∴ A hierarchical inflammasome assembly mechanism

❗: Discover
→ new mechanisms of (inflammasome formation)
❗: Uncover
→ the structural basis ← of hetero-oligomeric CARD-CARD interactions

Results

Cryo-EM structure determination of CARD8 and NLRP1 CARD filaments

Inflammasomes
→ leverage → supra-molecular filamentous structures
→ to nucleate (the polymerization of caspase-1)
→ increases → the local concentration of (its caspase domain) → to facilitate (dimerization & activation)

Expressed → these CTs
← in fusion with an N-terminal maltose-binding protein (MBP) tag
← separated by a linker cleavable ← by the human rhinovirus (HRV) 3C protease

A bulky tag
→ would disrupt → oligomerization
→ facilitate → purification of (monomeric proteins) → for controlled CT filament formation in vitro

MBP-fusion protein
→ formed → small oligomers

Purified → short filaments

Structure of the CARD8-CT filament

No UPP density
→ was observed
← in the CARD8-CT cryo-EM re-construction
← which only reveals → the central helical CARD filament

❗: UPA
→ does NOT follow → the CARD helical symmetry
→ is orderly associated ← with the central CARD filament

❓: Substracting out the central density
→ did not reveal → any new density

⁉: UPA & any residual MBP molecules
→ must be flexibly linked → to the core CARD filament
∴ Their stochastic position
← relative to the central CARD filament
→ caused them → to average out ← during (data processing)

❗: 1⃣ the UPA 2⃣ any residual un-cleaved MBP tag
→ appeared ← as noisy density ← surrounding the filaments

The CARD8-CARD filament
→ possesses → a one-start helical symmetry

Analysis
←of the three type of asymmetric interactions
← that are characteristic of death-fold filaments
→ revealed that → type I interaction is unusually small
→ for those in 1⃣ ASC-CARD 2⃣ caspase-1-CARD filaments

Detailed inspections
← of the three interfaces
→ revealed → many charge-charge pairs
→ as other hydrophilic interactions

Used → these structural insights
→ to design point mutations
← that would abolish CARD filament formation

Cleaved
→ 1⃣ recombinant wild-type 2⃣ mutant MBP-tagged CARD8-CARD proteins
← followed by (EM imaging) of (the negatively stained samples)
→ to assess → the impact of these mutations ← on filament formation

The wild-type protein
← which formed filaments ← when the bulky MBP tag
→ was removed → seven different point mutants
→ completely abolished (filament formation)

❗: A reconstitute HEK293T cell system
← which stably expressing 1⃣ caspase-1 2⃣ GSDMD
→ to assess whether these mutants → were able to signal in cells

Transfected → these cells ← with plasmid encoding
→ for WT or mutant
Analyzed → the supernatant
→ for lactate dehydrogenase (LDH) activity ← a hallmark of (pyroptotic cell death)
→ for the lysate by immunoblotting → for 1⃣ caspase-1 2⃣ GSDMD cleavage

Wild-type CARD8-UPA-CARD
→ induced → elevated LDH activity ← in the supernatant
→ exhibited → prominent 1⃣ caspase-1 2⃣ GSDMD cleavage

Five of six filament-deficient mutants
→ showed → background level LDH release
← with NO discernible 1⃣ caspase-1 2⃣ GSDMD cleavage

❗: expression of CARD8-CARD
→ did NOT lead → to inflammasome signaling
← under these same condition

Expressed CARD8-UPA-CARD
→ fused → to a C-termina mCherry tag ← in HEK293T cells
→ found that → the WT (← formed strong punctate structures) ← indicative of (filament formation)

All the CARD8-UPA-CARD
← which constructs ← with CARD interface mutations
→ showed diffuse distributions → consistent with defective (filament formation)

∴ These cellular data confirmed that
→ filamentous CARD-CARD interactions → are crucial
→ for CARD8-CT-mediated inflammasome signaling

UPA enhances NLRP1-CT and CARD8-CT signaling

CARD8-CARD construct itself
→ did NOT cause → cell death
CARD8-UPA-CARD → did
∴ ⁉ the presence of the UPA sub-domain
→ enhances → inflammasome formation

∴ Determine → the approximate minimum concentration
→ required → to form CARD8-CARD 🆚 CARD8-UPA-CARD filament in vitro

CARD8-CARD 🆚 CARD8-UPA-CARD
→ almost 100-fold changes

UPA sub-domain
→ significantly decreased → the concentration needed
→ to form 1⃣ CARD8 2⃣ NLRP1 filaments

1⃣ CARD8-UPA 2⃣ NLRP1-UPA
→ showed → aggregation ← under negative stain EM
∵ Ability → to promote filament formation

⁉: higher concentration ← of transfected plasmid
→ were required → for appreciable (cell signaling)
∵ CARD8-CARD formed filaments ← at very high concentration

❓: Increasing (the amount of transfected plasmid)
→ could lead → to 1⃣ inflammasome formation 2⃣ LDH release
∴ Raised → the plasmid amount

CARD8-CARD
→ significant induction ← of LDH release → ∴ cell death
→ only occurred ← when 2000ng of plasmid → was used

CARD8-UPA-CARD
→ caused → a similar level ← of LHD release

NLRP1-CARD
→ 2000 ng/transfection → led → to significant LDH release

NLRP1-UPA-CARD
→ 2000 ng/transfection → triggered similar levels ← of LDH release

∴ A 100-fold greater plasmid amount
→ is needed
→ to cause a similar level ← of CARD-mediated cell death

A front-to-back UPA oligomerization interface
→ required → for inflammasome signaling

UPA oligomerization
→ promotes → 1⃣ CARD8-CT & NLRP1-CT filament formation 2⃣ subsequent inflammasome signaling

Structure of the NLRP1-CT filament containing CARD dimers

No UPA density → was observed
∴ The UPA subdomain → is also flexibly connected
→ to the core NLRP1-CARD filament

❗: The NLRP1-CARD filament
→ has → one-start helical symmetry
← as other CARD filament

❓: the NLRP1-CARD filament
→ is composed ← of NLRP1-CARD dimers
→ is also reflected ← in the thicker dimension of NLRP1-CT filaments
← unlike other CARD filament

The inner NLRP1-CARD filament
→ is roughly equivalent
→ to other CARD filament structures

The dimerically ← related NLRP1-CARD subunits
→ form → the outer layer of the NLRP1-CARD filament

The NLRP1-CARD dimer
→ is mediated
← by reciprocal interactions

These regions ← of helices α5 and α6
→ are NOT involved ← in the inner core filament interaction
→ is mediated ← by the classical type I, II, & III CARD-CARD interactions

Charged & hydrophilic interactions
→ dominate → the interactions

Employed → the same reconstituted HEK293T cell sytem
→ which stably expressing 1⃣ caspase-1 2⃣ GSDMD
← which used from assessing CARD8 mutants in cells
→ to elucidate → the role of the observed NLRP1-CARD filament ← in signaling

Transfected → these cells
← with plasmids encoding ASC
Analyzed
→ the supernatant → for LDH activity
→ the lysate → for 1⃣ caspase-1 2⃣ GSDMD cleavage

NLRP1-CARD has a propensity for dimerization

Compared → the signaling activity
← of 1⃣ WT NLRP1-UPA-CARD 2⃣ NLRP1-UPA-CARD
← with several CARD dimerization mutant
→ to investigate → the function of NLRP1-CARD dimerization

Y1445A
→ comprised → LDH release
→ retained → caspase-1 processing
∴ Partial defectiveness

Other dimerization mutants
← e.g. 1⃣ M1457A 2⃣ W1460A 3⃣ E1461R
→ showed → no discernible impact
← on inflammasome signaling

The Y1445A mutant
→ abolished → NLRP1-UPA-CARD filament formation

The UPA
→ is required
→ for UPA-CARD inflammasome signaling
The dimer
→ promotes → assembly to a much lesser degree

❓: the CARD8-CT
→ fail → to induce CARD8-CARD dimerization?

Inspected → the crystal packing interactions
← in the previously determined 1⃣ MBP-fused NLRP1-CARD 2⃣ MBP-fused CARD8-CARD

NLRP1
→ all independent molecules
← in the crystallographic asymmetric unit
→ form → a symmetrical dimer ← in the crystal lattice
← which super-imposes well ← with the dimer observed in the filament

NLRP1-CARD
→ has an intrinsic propensity → to form dimers

The ASC type b surface nucleates caspase-1 polymerization

❗: how hetero-oligomeric CARD-CARD interactions
→ nucleate → caspase-1 filament assembly
→ to facilitate downstream signaling

ASC-CARD
→ uses → its type Ib, IIb, & IIIb surfaces
→ to interact with the caspase-1-CARD type Ia, IIa & IIIa surfaces
∵ these interactions → bury → a larger total surface area
← than that of the reverse interactions

Designed → mutations
← based on the ASC & caspase-1 filament structures
→ to prevent (filament elongation)
→ ∴ observed → a minimal unit of contact ← between ASC & caspase-1

Designed
→ 1⃣ the type IIa W169G mutant ← of ASC-CARD 2⃣ the type IIb G20K mutant ← of caspase-1-CARD
Connected them
← in one polypeptide chain ← using a GSS linker

Collected → a cryo-EM dataset
Found that → the octamer 2D classes → were quite detailed

∴ There are 1⃣ four ASC-CARD 2⃣ four caspase-1-CARD subunits

❗: This arrangement
→ requires → a tetramer layer ← of ASC-CARD
→ to interact with a tetramer layer of caspase-1-CARD
∴ A hierarchical assembly of these (death domains)

Extensive type I & II interfaces
→ dominate → the ASC-CARD-caspase-1-CARD assembly
← with calculated buried surface areas

Detailed inspection ← of these interfaces
→ revealed → several clear interactions
← between 1⃣ ASC-CARD 2⃣ caspase-1-CARD

Predicted modes of interactions between NLRP1-CARD and ASC-CARD and between CARD8-CARD and caspase-1-CARD

Most CARD-containing inflammasome proteins
← e.g. 1⃣ NLRP1 2⃣ NLRC4
→ amplify signaling → through the adapter protein ASC

CARD8
→ is the only known inflammasome sensor
← which cannot engage ASC
← which instead exclusively binds caspase-1

These differences
← in downstream signaling components
→ lead → to different biological outcomes
∴ The signal amplification & its differential expression ← across cell types

The structures ← of 1⃣ CARD8-CARD 2⃣ NLRP1-CARD filaments
→ predict → the modes of their CARD-CARD interactions
∴ A uni-directional assembly → for several other death domain super-family members

Investigated
→ the cross-sectional surface charges
← of 1⃣ CARD8 2⃣ NLRP1 3⃣ ASC 4⃣ caspase-1 filament

Generated
→ a CARD8-CARD--caspase-1-CARD hypothetical complexs
← by fitting a layer of caspase-1-CARD tetramer
→ to the CARD8-CARD filament structure
← by rigid-body fitting

The interfacial side chains
← of 1⃣ CARD8-CARD 2⃣ caspase-1-CARD
→ were NOT 1⃣ adjusted 2⃣ energy minimized
∴ These model should be taken ← as highly suggestive

The charge complementarity
← between 1⃣ CARD8-CARD 2⃣ caspase-1-CARD
→ is apparent

Both interfaces
→ display largely negative charges
→ should repel each other

Charge complementary
← with negative & positive patches
→ near the center of the filament cross-sections

Analyzed → the modeled interface (type by type)

❗: The structural analysis
→ confirmed → favorable interactions

ASC
→ uses its type b surface
→ to recruit caspase-1
∴ A hierarchical inflammasome formation
← which proceeds through an ordered manner
← with potential signal amplification

This ordered assembly
→ is made
← by the use of two opposing ASC filament surfaces

Oligomerized ASC-CARD
→ could nucleate → the associated ASC-PYD
→ to form helical filaments
∴ recruit → more ASC ← through PYD-PYD interactions
∴ allow the recruited ASC → to template more caspase-1-cCARD filament formation

---