Biological context

Inflammasomes are multiprotein oligomers that are of fundamental importance for the innate immune system. The formation of these oligomers is triggered by the recognition of pathogen- and danger-associated molecular patterns (PAMPs and DAMPs) by specific, germ-line encoded pattern-recognition receptors (PRRs), and eventually leads to an inflammatory cascade. This cascade starts with the activation of caspase-1 and the maturation of the inflammatory cytokines Interleukin 1β (IL-1β) and Interleukin 18 (IL-18), leading eventually to cell death. It has been shown that the structures of inflammasome complexes are crucial for the efficiency of the immune response (Schroder and Tschopp 2010); moreover, malfunctions of such oligomers are related to human disease, such as cancer and autoimmune syndromes (Schroder and Tschopp 2010; Hu et al. 2013).

The signaling pathway from the PRRs to the activation of caspase-1 is mediated in many cases by the bipartite adaptor apoptosis-associated speck-like protein containing a CARD (ASC), which consists of an N-terminal PYD and a C-terminal CARD. In ASC-dependent inflammasomes, the PYD oligomerizes upon interaction with the receptor to form the so-called ASC speck, and the C-terminal CARD subsequently recruits the inflammatory caspase. The large ASC filament thus forms the structural core of the ASC-inflammasome (Masumoto et al. 1999).

The monomer structures of the human ASC-PYD (de Alba 2009), and full-length ASC (Liepinsh et al. 2003) have previously been obtained by solution-state NMR spectroscopy, and the interaction mode of PYDs in the inflammasome assembly was addressed (Vajjhala et al. 2012; Srimathi et al. 2008; Vajjhala et al. 2014; Moriya et al. 2005). Recently, the structure of the filaments of the human ASC-PYD has been solved at ∼3.8 Å resolution using cryo-electron microscopy (Lu et al. 2014). In this work we present the chemical-shift assignment for the filaments of the mouse PYD of ASC, as the basis for the structural study of the filaments. We show that the presence or the absence of the CARD does not influence the structure of the filaments formed by the PYD.

Methods and experiments

Expression, purification and filament formation

The full length mouse ASC protein (residues 1–193) and the ASC-PYD domain only (residues 1–91) was produced using a PET28a vector under the control of a T7 promoter. A GSGSLE linker and a His6-tag were added to the C-terminus of both protein constructs. Proteins were expressed in BL21(DE3) E. coli strains by growing the cultures at 37 °C to an OD600 of 0.8, inducing with 1 mM IPTG and harvesting after 4 h. [U-15N,13C]-labeling was obtained using 13C-glucose and 15NH4Cl. The cells were harvested by centrifugation and the pellet was resuspended in 50 mM phosphate buffer pH = 7.5, 300 mM NaCl, 0.1 mM Complete protease inhibitor (Roche). The obtained cells were incubated for 1 h at room temperature with DNase I and then sonicated on ice and centrifuged at 13,000 rpm at 4 °C for 30 min. The pellet was resuspended in 50 mM phosphate buffer pH = 7.5, 300 mM NaCl, 6 M guanidinium hydrochloride and centrifuged at 13,000 rpm at 4 °C for 30 min. The supernatant was incubated for 2 h at room temperature with pre-equilibrated Ni-NTA affinity resin (Thermo Scientific) and than passed through a plastic body column for gravity flow purification. The column was washed with 20 column volumes of resuspension buffer containing 20 mM imidazole and eluted with 3 column volumes of the same buffer with 500 mM imidazole. All these purification steps were carried out at 4 °C. Filaments were prepared by slow dilution into aqueous solution. The elution fraction from the NiNTA column was dialyzed overnight against 25 mM phosphate buffer pH = 7.5, 100 mM NaCl. The solution was then centrifuged at 13,000 rpm at 4 °C for 30 min yielding a gel-like pellet of ASC filaments that was resuspended in water for a wash step and stored at 4 °C. This same protocol also yielded samples for EM studies. Details thereof are discussed in a forthcoming publication.

NMR sample preparation

The filaments so obtained were packed into 3.2 mm ZrO2 rotors (Bruker Biospin) by ultracentrifugation using a homemade filling device (Böckmann et al. 2009) in a SW41-T1 swing-out rotor in an optima L90-K ultracentrifuge (Beckmann). The sample was spun for 12 h, setting the spinning to 25,000 rpm at 4 °C. The drive tip of the rotor was finally sealed with epoxy glue (Araldit® blue) in order to prevent the dehydration of the samples.

Solid-state NMR spectroscopy

3D experiments were performed for the assignment, namely NCACB, NCACX, NCOCX and CANCO, as well as the relayed N(CO)CACB and CAN(CO)CA experiments for the assignment of the backbone and Cβ atoms of the protein (Schuetz et al. 2010) 2D NCA and DARR were used to check assignment consistency. Additional 3D CCC, N(CA)CBCX and 2D DARR were performed to continue the assignment of the side chains. All 3D spectra were recorded for 2–6 days each (Table 1). 20 ms DARR spectra on ASC-PYD and ASC full-length filaments were recorded for comparison.

Table 1 Experimental parameters for the solid state NMR spectra used for the sequential assignment of the ASC-PYD
Full size table

All spectra were measured on a Bruker Avance II+ 850 MHz operating at a static magnetic field of 20 T. Experiments were performed with magic angle spinning (MAS) at 17 kHz with a Bruker 3.2 mm probe. The sample temperature was 12 °C, as determined from the water resonance (Gottlieb et al. 1997). Processing was done using TopSpin 3.1 (Bruker Biospin) by zero filling to no less than double the number of acquired points. The time domain was apodized using a squared cosine function (SSB 2.2–2.6). The spectra were analyzed and annotated with ccpNMR analysis 2.3 software (Fogh et al. 2002; Vranken et al. 2005; Stevens et al. 2011).

Secondary-structure analysis

Secondary chemical shifts were obtained by subtracting the random-coil shifts, as tabulated in (Wang and Jardetzky 2002). The analysis was done using the program TALOS+ (Shen et al. 2009).

Assignment and data deposition

The [13C–13C] and [15N–13C] correlation NMR spectra of the mouse ASC PYD filaments feature narrow lines (line width of cross peaks on the order of 0.5 ppm), which is a clear indication of a homogeneous sample. The filaments showed no polymorphism, as shown in Fig. 1: all the peaks present in the 20 ms DARR (Fig. 1a) and in the NCA (Fig. 1b) were assigned to a single nucleus, with the exception of two additional weak threonine signals present in the DARR. Most likely, these two signals arise from Thr 87 and Thr 88, but no nitrogen resonance and no connectivity information could be obtained from the different spectra. Figure 2 compares the 20 ms DARR (Fig. 2a) and the NCA (Fig. 2b) measured on the filaments of the isolated PYD (blue) with the full length ASC (red). The very good coincidence of the resonances in these spectra indicates that the rigid part of the filaments is composed only by the PYD, while the CARD shows dynamic behavior. Moreover, the presence of the CARD does not influence the structure of the PYD in the filaments. All measurements for sequential assignments were thus performed on samples of filaments of the isolated ASC PYD. Using 3D methods (Schuetz et al. 2010; Habenstein et al. 2011), all backbone atoms from residues 3–85 could be assigned. The relayed 3D experiments where no evolution of the C′ was recorded [e.g. N(CO)CACB] were fundamental to reduce the ambiguities in some stretches of the protein because of the larger shift dispersion of the Cβ compared to C′. For the CO-CA transfer, whose efficiency is important for these experiments, we have used the efficient band selective homonuclear CO-CA cross polarization (Chevelkov et al. 2013). In Fig. 3, a representative strip plot of the sequential assignment is shown, using a combination of the relayed 3D experiments with an NCACO, in which equally side-chain resonances can be observed. A total of about 80 % of all side-chain carbon nuclei were assigned with additional 2D and 3D spectra, including CCC (Fig. 4). The non-assigned nuclei were either not visible in the spectra due to local dynamics (as for the side chains of 5 Arg, 25 Met Cε, 70 Met Cε) or not assignable due to spectral overlap (as for 6, 51, 54 Asp Cε). The chemical shifts have been deposited in the BMRB database under the accession number 26550. The first three and the last six residues were only visible in some of the dipolar-transfer-based NMR spectra and could not be sequentially assigned with certainty. We performed an INEPT-type [15N–1H] spectrum of the ASC PYD filaments to investigate the flexibility of these stretches, but it was devoid of peaks (data not shown). We concluded that these residues experience dynamics on the intermediate time scale, and are thus line-broadened below the detection limit using either solution- or solid-state NMR methods. In Fig. 5, an overview of the assignment is shown. The secondary chemical shifts derived from the chemical shifts reported here using a TALOS+ analysis is shown in Fig. 6, and points to the presence of six α-helices, at positions 3–14, 17–29, 41–46, 49–59, 62–76, and 80–84. These locations are similar to those observed in the monomeric human ASC PYD using solution-state NMR, where helices were defined including residues 3–14, 17–26, 41–46, 49–59, 62–76, 80–89 (Liepinsh et al. 2003).

Fig. 1
figure 1

Solid-state NMR spectra of uniformly 13C, 15N labeled mouse ASC-PYD filaments acquired at a magnetic field of 20 T and MAS at a frequency of 17 kHz. a 2D 13C–13C DARR spectrum with a mixing time of 20 ms. b NCA spectrum. The sequence-specific assignments indicated were derived from the sequential backbone walk in 3D spectra as shown in Fig. 3

Full size image
Fig. 2
figure 2

Comparison of 2D 13C–13C DARR spectrum with a mixing time of 20 ms (a) and NCA (b) measured on uniformly 13C–15N labeled filaments of PYD alone (red) and full-length ASC (blue)

Full size image
Fig. 3
figure 3

Strips extracted from 3D NCACO, N(CO)CACB and CAN(CO)CA spectra of ASC-PYD filaments. Vertical lines where added to follow the backbone walk starting at the first visible residue of the N-terminus of the protein

Full size image
Fig. 4
figure 4

Strips extracted from the 3D CCC spectrum of ASC-PYD filaments, used for the assignment of the sidechains. In a CCC spectrum, polarization transfers are selected to mainly display Cα resonances in the first dimension, Cβ in the second and the remaining side chains in the third dimension. In a aromatic sidechains for Tyr 59 are shown, in b sidechains for different Ile and in c for Leu 32

Full size image
Fig. 5
figure 5

Assignment graph of ASC-PYD filaments created using the CPPN software (Fogh et al. 2002; Vranken et al. 2005; Stevens et al. 2011). Black dots indicate assigned spins, while the unassigned ones are displayed in white

Full size image
Fig. 6
figure 6

Secondary chemical shifts analysis of labeled filaments of ASC-PYD. For glycine ∆δCα is plotted and the corresponding bars are plotted in gray. Four positive secondary shifts in a row are indicative of α-helix structure. Blue squares are drawn for TALOS+ predicted α-helix regions. Only residues without warning and with small dispersion of predictions were included

Full size image

The secondary-structure elements determined by the secondary chemical shifts will be used in a forthcoming publication, in combination with high-resolution cryo electronmicroscopy data, to determine a de-novo structure of the ASC PYD filament.