A motif within the armadillo repeat of Parkinson's-linked LRRK2 interacts with FADD to hijack the extrinsic death pathway

Abstract

In experimental models, both in vivo and cellular, over-expression of Parkinson's linked mutant leucine-rich repeat kinase 2 (LRRK2) is sufficient to induce neuronal death. While several cell death associated proteins have been linked to LRRK2, either as protein interactors or as putative substrates, characterization of the neuronal death cascade remains elusive. In this study, we have mapped for the first time the domain within LRRK2 that mediates the interaction with FADD, thereby activating the molecular machinery of the extrinsic death pathway. Using homology modeling and molecular docking approaches, we have identified a critical motif within the N-terminal armadillo repeat region of LRRK2. Moreover, we show that co-expression of fragments of LRRK2 that contain the FADD binding motif, or deletion of this motif itself, blocks the interaction with FADD, and is neuroprotective. We further demonstrate that downstream of FADD, the mitochondrial proteins Bid and Bax are recruited to the death cascade and are necessary for neuronal death. Our work identifies multiple novel points within neuronal death signaling pathways that could potentially be targeted by candidate therapeutic strategies and highlight how the extrinsic pathway can be activated intracellularly in a pathogenic context.

Figure 1

Genetic deletion of Bax prevents mutant LRRK2 induced neuronal death. Primary embryonic cortical neurons are derived from WT mice or mice lacking the pro-apoptotic protein Bax, and transiently transfected with GFP-tagged WT or mutant LRRK2 (R1441C, G2019S). (a) A representative image showing typical morphological features of apoptotic neurons expressing mutant LRRK2; that are positive for active caspase-3 (casp3) and display condensed and fragmented nuclei. (b) Quantification of apoptotic neuronal death (apoptotic nuclear profiles, “apo”; and active caspase-3 positive neurons, “casp3”) from cultures depicted in (a); **p < 0.01 in comparison to WT transfected neurons. (c) Primary neurons were prepared from WT mice or mice deficient in Bax, and treated with the topoisomerase I inhibitor camptothecin (“cam”, 5 μM, 18 hr), or doxorubicin (“doxo”, 10 ng/ml, 36 hr). Cells were fixed and processed for apoptotic nuclear counts. (d) Primary cortical neurons were prepared from WT mice or mice deficient in Bax, and transiently transfected with WT, R1441C, or G2019S LRRK2. As before, the cultures were fixed and the percentage of LRRK2-positive neurons that displayed apoptotic nuclear features was determined. The absence of Bax significantly protected neurons from death induced by mutant LRRK2. ***p < 0.001 compared to WT neurons expressing R1441C- or G2019S-LRRK2. Each data point represents the mean +/− SEM from 3–4 independent transfections from a representative culture. Cultures were repeated at least 3 times with similar differences.

Figure 2

Pharmacological inhibition of Bax and Bid block apoptotic death induced by mutant LRRK2. WT cortical neurons transiently expressing WT, R1441C, G2019S, or I2020T LRRK2 (with an EGFP reporter) were vehicle treated, or treated with a pan caspase inhibitor (VAD), or peptide inhibitors of Bax (V5) or Bid (BI-6C9). (a) A representative confocal image of G2019S mutant LRRK2 expressing neurons, treated with vehicle (DMSO), VAD, V5, or BI-6C9. The percentage of neurons positive for the EGFP reporter and anti-LRRK2 (clone c41–2) exceeded 85%. The stained neurons were counted in a blinded fashion, and the percentage of GFP-positive neurons with apoptotic nuclear features was determined. Inhibition of caspase activation, as well as translocation of Bax or truncated Bid to the mitochondria, protects neurons from apoptotic death induced by different mutant forms of LRRK2. *p < 0.05 compared to untreated R1441C-LRRK2 expressing neurons; **p < 0.01 compared to untreated G2019S or I2020T-LRRK2 expressing neurons. Each data point represents the mean +/− SEM from 3–4 independent transfections from a representative culture. Cultures were repeated at least 3 times with similar differences. Asterisks indicate LRRK2-expressing neurons.

Figure 3

The N-terminal region upstream of the LRRK2 leucine-rich repeat interacts with FADD. (a) A schematic of the domain structure of LRRK2 with amino acid boundaries used in this study to map the interacting domain with FADD (lower schematic). (b) V5-tagged FADD was co-expressed in HEK293T cells with Flag-tagged LRRK1, or Flag-tagged constructs of LRRK2, encoding the full-length protein, or each of the discrete domains as illustrated in (a). Cell lysates were incubated with anti-Flag resin and the inputs and eluates were separated by 12% SDS-PAGE, followed by probing the membranes with anti-V5, anti-Flag, or anti-β-actin. The upper blots show the immunoprecipitation of Flag-LRRK1 or Flag-LRRK2 and probed for anti-Flag, and the co-precipitation of V5-FADD is shown in the middle panel (IP eluate) probed with anti-V5. FADD interacts with full-length LRRK2, but not the related LRRK1, and binds only the N-terminal domain upstream of the LRR. The left blot (in panel [b]) shows a longer exposure of the region highlighted by the box. Identical amounts of protein from the input lysate from each sample was separated by 12% SDS-PAGE, and the membranes probed with anti-Flag, anti-V5, and anti β-actin (c). In the blot shown in (d), identical lysates from cells expressing Flag-LRRK1 or Flag-LRRK2 were separated by 8% SDS-PAGE, to allow better separation of the LRRK1 and LRRK2 proteins.

Figure 4

The 3D homology model of the dimeric LRRK2 ARM domain. (a) The armadillo repeats of the two LRRK2 molecules establish a concave inner conformation to each monomer. In this model of the isolated ARM repeat domain, the sites capable of dimerization are lined with residues that establish a hydrophobic core. Shown are each of the modeled LRRK2 monomers (presented in blue and magenta ribbon) of the LRRK2 ARM region. Each structure bears a specific α-helical motif that is exposed on the molecule located near the C-terminus of the ARM repeat region. One such motif is highlighted in green. (b) When viewed from the side, the two monomers display a supercoiled dimeric-like formation that is typical of armadillo repeat containing proteins.

Figure 5

Molecular docking of LRRK2 ARM to FADD-DD. The model of dimeric LRRK2 ARM was used to dock the FADD DD crystal structure. The docking was performed iteratively placing each FADD DD structure onto the LRRK2 ARM model in sequence. (a) The FADD DD crystal structure (brown ribbon) was docked onto the LRRK2 dimer model (shown in blue and magenta ribbon, as before). (b) A second copy of the FADD DD crystal structure (silver ribbon) was docked to the molecular complex established in the previous step (a). The interaction was achieved via a network of antiparallel α-helical bundles (shown in green ribbon representation). (c) The network of molecular interactions between the specific residues involved is highlighted (inset). (d) The association between LRRK2 and FADD DD is stabilized via a mixture of ionic/hydrogen bonds, π-stacking and hydrophobic interactions. The amino acid notations are as follows: A, refers to the LRRK2 ARM sequence; and B, C refer to each FADD DD molecule.

Figure 6

Molecular docking of the full LRRK2-FADD complex. (a) The dimeric model of LRRK2 ARM was used to dock two molecules of FADD DD on each end. By 180° rotation of the horizontal axis of the dimer interface (b), it is apparent that the axis of each group of triple-helix bundles is perpendicular to the other. Note that the established interacting α-helical bundles have been highlighted in red and green colored ribbon.

Figure 7

Deletion of predicted FADD binding domain in LRRK2 ARM prevents recruitment of LRRK2 to FADD complexes. Full-length mutant R1441C-LRRK2, or R1441C-LRRK2 lacking the predicted FADD binding domain (specifically, the residues within this motif with the highest identity, 538-547), was co-transfected with V5-FADD in HEK293T cells. Following 48 h of expression cells were fixed and immunostained for Flag (LRRK2) and V5 (FADD), together with DAPI. Over-expression of FADD leads to its localization to so-called Death Effector Filaments (DEFs) that subsequently recruits LRRK2. Deletion of the motif predicted to mediate binding to FADD (ΔFBD; FADD-binding domain) prevents this recruitment, leaving mutant LRRK2 diffusely localized in the cytoplasm.

Figure 8

Quantification of LRRK2 recruitment to FADD DEFs. The percentage of LRRK2 filamentous structures co-localized with FADD DEFs was determined in a blinded fashion for each pair of full-length or deletion mutant LRRK2 constructs: R1441C, G2019S, and I2020T. **p < 0.01 compared to full-length mutant LRRK2; ***p, 0.001 compared to full-length mutant LRRK2. Each data point represents the mean +/− SEM from 3–4 independent transfections.

Figure 9

Deletion of predicted FADD binding domain in LRRK2 ARM region is neuroprotective. Primary rat embryonic cortical neurons were transiently transfected with full-length mutant, or its corresponding deletion pair (ΔFBD), together with a construct encoding an EGFP reporter at a ratio of 4:1. Following 72 h of expression, neurons were fixed and processed for anti-LRRK2/EGFP immunofluorescence with DAPI nuclear stain (a). Please see Fig. 2 and Supplementary Fig. 11 for representative images of full-length mutant LRRK2. Panel (b) shows the relative expression levels of full-length or deletion constructs. HEK293T cells were transiently transfected with full-length mutant LRRK2, or mutant LRRK2 (ΔFBD) lacking the FADD binding domain. Cell extracts were separated by SDS-PAGE and the membranes probed with anti-LRRK2 (c41-2) and β-actin, as a loading control. In (c), the percentage of EGFP-positive neurons displaying apoptotic nuclear morphology was determined by a rater blinded to the experimental conditions. **p < 0.01 compared to full-length mutant LRRK2; ***p < 0.001 compared to full-length mutant LRRK2. Each data point represents the mean +/− SEM from 3-4 independent transfections. Cultures were repeated at least 3 times with similar differences.

Figure 10

Fragments of LRRK2 containing FADD binding motif are neuroprotective. (a) HEK293T cells were co-transfected with V5-FADD and Flag-tagged full-length WT or mutant LRRK2 (I2020T); or fragments of the N-terminal domain of LRRK2 comprised of amino acids 1-500 (NT500), 1-550 (NT550), 1-575 (NT575), or 1-969 (NT969). Cell extracts were subjected to anti-Flag immunoprecipitation and the eluates were probed for the presence of co-precipitating FADD. While a fragment of the LRRK2 N-terminal domain containing amino acids 1-500 (NT500) was unable to bind FADD, increasing the length to 550 or 575 amino acids (NT500, NT575) restored the interaction with FADD, suggesting that the FADD binding motif is located within residues 500-575 of the LRRK2 N-terminal domain. (b) Primary cortical neurons were transiently co-transfected with full-length WT or R1441C, G2019S, or I2020T mutant LRRK2 together with the different N-terminal fragments of LRRK2 depicted in the schematic. Fixed neurons were then assessed for the presence of apoptotic nuclear features. Co-expression of the fragments of LRRK2 that interact with FADD blocked the apoptotic death of cortical neurons induced by full-length G2019S mutant LRRK2. **p < 0.01 compared to neurons transfected with R1441C or I2020T LRRK2 and pcDNA control plasmid; ***p < 0.001 compared to neurons transfected with G2019S-LRRK2 and pcDNA control plasmid. Each data point represents the mean +/− SEM from 3–4 independent transfections. Cultures were repeated at least 3 times with similar differences.

Figure 11

Fragments of LRRK2 ARM region are neuroprotective. Primary cortical neurons were transiently co-transfected with full-length untagged WT or G2019S mutant LRRK2 together with different Flag-tagged N-terminal fragments of LRRK2 depicted in the schematic shown in Fig. 10b. Fixed neurons were then processed for anti-Flag immunostaining to label the LRRK2 ARM fragment (NT500, NT575, or NT969), together with anti-LRRK2 (clone c41-2) to label full-length LRRK2. A representative image showing neurons expressing G2019S-LRRK2 together with the different ARM fragments is presented. Scale bar is 10 μm.

Figure 12

Proposed model of FADD/Bax-dependent neuronal death induced by mutant LRRK2. Dimeric mutant LRRK2, binding FADD through its ARM repeat region located within the N-terminal domain, recruits and activates caspase-8. We show here that the intrinsic mitochondrial apoptotic machinery is in turn activated through the pro-apoptotic Bcl-2 proteins, Bid and Bax.