α-Synuclein fibril-specific nanobody reduces prion-like α-synuclein spreading in mice

Pathogenic α-synuclein (α-syn) is a prion-like protein that drives the pathogenesis of Lewy Body Dementia (LBD) and Parkinson’s Disease (PD). To target pathogenic α-syn preformed fibrils (PFF), here we designed extracellular disulfide bond-free synthetic nanobody libraries in yeast. Following selection, we identified a nanobody, PFFNB2, that can specifically recognize α-syn PFF over α-syn monomers. PFFNB2 cannot inhibit the aggregation of α-syn monomer, but can significantly dissociate α-syn fibrils. Furthermore, adeno-associated virus (AAV)-encoding EGFP fused to PFFNB2 (AAV-EGFP-PFFNB2) can inhibit PFF-induced α-syn serine 129 phosphorylation (pS129) in mouse primary cortical neurons, and prevent α-syn pathology spreading to the cortex in the transgenic mice expressing human wild type (WT) α-syn by intrastriatal-PFF injection. The pS129 immunoreactivity is negatively correlated with the expression of AAV-EGFP-PFFNB2. In conclusion, PFFNB2 holds a promise for mechanistic exploration and therapeutic development in α-syn-related pathogenesis.

Supplementary figure 2. Testing the binding of the disulfide bond-free GFP nanobody on the yeast surface. a) Schematics of the binding assay on the yeast surface. Anti-GFP nanobody (GFPNB) with and without disulfide bond was expressed on the yeast surface following Aga2p. The yeast cells were incubated with EGFP-mCherry fusion protein. b) FACS analysis of yeast cells expressing GFPNB-FLAG on the yeast surface incubated with EGFP-mCherry fusion protein. GFPNB without disulfide bond (C22L, C96A mutant) retains its binding to EGFP. Negative control (-GFPNB) is yeast cells without induction therefore no GFPNB expression.
Supplementary figure 3. Amino acid sequence schematics of the nanobody library design and FACS analysis of the nanobody selection. a) Nanobody library construction based on a published protocol 34 , except that the conserved cysteine residues were mutated in our library to remove the conservative disulfide bond under oxidizing conditions. C22L and C95A mutations are indicated by orange letters. The constant regions of the nanobodies were determined from the consensus sequences of the VHH from the llama gene IGHV1S1-IGHV1S1S5 and are shown in grey with only one amino acid in each position 34 . The CDR1, 2, and 3 are highlighted in blue, green, and red. X indicates site-saturated randomization with 20 amino acids. Some positions in grey have multiple amino acids in the same position, indicating randomization with those amino acids. Additionally, the CDR3 was constructed with 3 different lengths with 7, 11, or 15 amino acids randomized (red). b) Gating strategy to analyze single yeast cells. First, cells were plotted by FSC-A and SSC-A, and a gate P1 was drawn to include almost all the cells. Cells from P1 were then plotted by FSC-W and FSC-H and a gate between 60 -110 FSC-W and 0 -255 FSC-H gave population P2. Cells from P2 were then plotted by SSC-W and SSC-H and a gate between 60 -105 SSC-W and 0 -195 SSC-H gave population P3. Cells from population P3 were analyzed to show FLAG signal in the x-axis (640 nm laser and 670/14 emission filter) and α-syn signal in the y-axis (561 nm laser and 586/15 emission filter). c) FACS analysis of the rounds of nanobody selection against α-syn PFF. The red trapezoid indicates the selection gate for FACS. After one round of MACS followed by 3 more rounds of FACS, a false positive nanobody population showed up with high fluorescence signal even in the absence of α-syn PFF incubation (post 4 th round). Removal of the false-positive clones on the 5 th round sorting using MACS was attempted but not successful (blue arrow, post 5 th round). This is possible because some low-expressing yeast cells can escape the negative selections using MACS, but will show up in the next round after reamplification. Eventually, we used FACS to draw a tight gate to select only the true positive population, avoiding the false positive population (red triangle, post 5 th round).
Supplementary figure 4. FACS analysis of the 28 nanobody clones selected against α-syn PFF. Yeast cells expressing 28 different nanobody clones were incubated with α-syn PFF, monomers, or just buffer (control). Then cells were then labelled with mouse anti-α-syn antibody and anti-mouse IgG antibody conjugated to AlexaFluor 568. The numbers in the upper right corner of Q2 indicate the ratio of Q2/Q4 population. All 28 clones showed selective binding to α-syn PFF over α-syn monomers.
Supplementary figure 5. Analysis of the CDR3 of the 28 PFFNB clones. a) CDR3 length analysis of the 28 PFFNB clones. 47% of the nanobody clones consist of a CDR3 with 7 amino acids randomized, 32% with 11 amino acids, and 21% with 15 amino acids randomized. b) The CDR3 of the selected PFFNBs have a net positive charge and are rich in hydrophobic residues. Source data are provided as a Source Data file.
Supplementary figure 6. SDS-PAGE analysis of the PFFNBs expression and purification in E. coli. a) Protein expression in E. coli BL21. SDS-PAGE analysis of crude HisTag-MBP-PFFNBs proteins. The expected protein size is ~60 kDa. The majority of the protein at ~ 60 kDa (green arrow) was retained in the cell pellet while the extracted protein was at ~ 50 kDa (red arrow), possibly due to early termination of translation or truncation. Positive control protein HisTag-MBP-GFPNB(C22L, C96A) (positive control) appeared at the correct molecular weight (blue arrow). b) SDS-PAGE of HisTag-MBP-PFFNB or GFPNB expressed in E. coli BL21(C14) with co-expressed chaperones. All the crude PFFNB protein extract appeared at ~ 60 kDa, the correct molecular weight. A major band at ~60 kDa was also detected in the pellet. c) SDS-PAGE of purified HisTag-MBP, HisTag-MBP-PFFNB2, and HisTag-MBP-NbSyn87 after size exclusion chromatography, which were used in Fig. 2 and Supplementary Fig. 7. All experiments were replicated twice with similar results. Source data are provided as a Source Data file. Wells were coated with 3 ng/μL of α-syn PFF, and then titrated with 3. 3, 33.3, 66.7, 133.3, 266.7, 666.7, and 1333.3 nM of MBP-PFFNB2 or MBP alone. Three data points were collected for each concentration and shown as mean ± SEM. b) ELISA analysis of anti-α-syn mAb binding to α-syn monomers and PFF. Wells were coated with 3 ng/μL of α-syn PFF or monomers, and then titrated with 1. 3, 5.3, 13.3, 53.5, 106.6, and 213.3 nM of anti-α-syn mAb. Three data points were collected for each concentration and shown as mean ± SEM. c) ELISA analysis of MBP-NbSyn87 binding to α-syn monomers and PFF. Wells were coated with 3 ng/μL of α-syn PFF or monomers, and then titrated with 3. 3, 33.3, 66.7, 133.3, 266.7, 666.7, and 1333.3 nM of MBP-NbSyn87. Two data points were collected for each concentration. d) ELISA analysis of MBP-PFFNB2 binding to recombinant human wildtype (WT) α-syn PFF (WT-PFF) and α-syn(A53T) PFF (A53T-PFF). Wells were coated with 3 ng/μL of wild-type α-syn PFF or α-syn(A53T) PFF then titrated with 3. 3, 33.3, 66.7, 133.3, 266.7, 666.7, and 1333.3 nM of MBP-PFFNB2. Three data points were collected for each concentration and shown as mean ± SEM. e) ELISA analysis of PFFNB2 and PFFNB2 (L22C, A95C) binding to α-syn PFF. Wells were coated with 3 ng/μL of α-syn PFF or monomers, and then titrated with 3. 3, 33.3, 66.7, 133.3, 266.7, 666.7, and 1333.3 nM of MBP-PFFNB2 or MBP-PFFNB2 (L22C, A95C). Three data points were collected for each concentration and shown as mean ± SEM. f) Immunoblot analysis of the soluble and insoluble fractions of mice brain lysates with and without α-syn pathology, and Snca knock-out (KO) mice brain lysate, with indicated antibodies. All experiments above was replicated once with similar results. Source data are provided as a Source Data file.
Supplementary figure 8. The ThT assay for α-syn aggregation assay with PFFNB2. Effect of MBP-PFFNB2 or MBP alone on α-syn (2 mg/mL) aggregation with the ThT assay. Quantification data are the means ± SEM, n = 3 independent experiments, P values were determined by two-sided Student's t-test. (MBP + α-syn vs. MBP-PFFNB2 + α-syn P = 0.9404). ns, not significant. Source data are provided as a Source Data file.
Supplementary figure 9. EGFP-PFFNB2 was expressed after PFF treatment and AAVs encoding EGFP-PFFNB2 did not cause neurotoxicity. a) WT mouse primary cortical neurons transduced with AAVs encoding EGFP (AAV-EGFP) or AAV-EGFP-PFFNB2. Cells were analyzed on indicated days in vitro (DIV) for EGFP expression. Red arrows indicate the day of AAV transduction and α-syn PFF administration, scale bar, 50 μm. b) Quantification of EGFP positive cells per 100 cells. n = 3 independent experiments. Quantification data are the means ± SEM. c) Primary cortical neurons were treated with AAVs encoding EGFP or EGFP-PFFNB2 on 5 DIV, and were assessed for neurotoxicity (NeuN immunostaining) 10 days after AAV transduction. scale bar, 50 μm. d) Quantification of NeuN positive cells (panel c) were counted per area and normalized to control n = 6 (from 3 independent experiments performed in duplicate). Quantification data are the means ± SEM, statistical significance was calculated by two-sided Student's t-test. (AAV-EGFP vs. AAV-EGFP-PFFNB2 P = 0.9392). ns, not significant. Source data are provided as a Source Data file.