Fig 2: A genome-wide CRISPR activation screen identifies restriction factors for influenza B virus.a Schematic of the CRISPR activation screening approach used. sgRNA = single guide RNA. b sgRNA read counts detected for each round of infection by B/Yamagata/16/1988 (Yam/88) during the screen. Each point represents a unique sgRNA for each independent screen replicate and sgRNAs not detected are omitted from the graph. c Ranked scores for upregulated genes after three rounds of Yam/88 infection. The top six ranked genes are labeled. d Expression of the indicated genes in A549 dCas9-VP64 cells after activation with the most enriched sgRNAs from the screen or a non-targeting control sgRNA, measured using qRT-PCR (mean with SEM, N = 4 experiments, B4GALNT2 *p = 0.0286, B3GAT1 *p = 0.0286, TGM2 *p = 0.0286, ANXA11 p = 0.8857). ND = not detected. e Percent of cells infected with Yam/88 PB1-mNeon (MOI = 0.6, 24 HPI, single cycle) in the indicated sgRNA-activated cell lines, measured using flow cytometry (mean with SEM, N = 4 experiments, B4GALNT2 *p = 0.0286, B3GAT1 *p = 0.0286, NXPE4 *p = 0.0286, TGM2 p = 0.6857, ANXA11 p = 0.8857, GPHA2 p = 0.6857). Statistical comparisons were all calculated with respect to the non-targeting control. All statistical analyses were performed using a two-tailed Mann-Whitney U test. * indicates p < 0.05, ns = not significant, nt = not tested. Source data are provided as a Source Data file.

Fig 3: Neurite branch points of motor neurons treated with different concentrations of HNK1 mimetics as measured with automated InCell analysis. Cultures of motor neurons were treated for 24 h with different concentrations of HNK1 mimetics or as positive control with HNK1 mimetic peptide. DMSO (0.2%) served as vehicle control (0). Neurite branch points of β-tubulin class III immunostained neurons were counted by the InCell Analyzer. Bar diagrams show the average number of branch points per neurite (n ≥ 2000 neurons from three independent experiments +SEM). Asterisks show differences in branch points between HNK1 mimetics and the vehicle control (one-way ANOVA; Fisher’s PLSD test, * p < 0.05, ** p < 0.01, *** p < 0.001).

Fig 4: B3GAT1 can be safely expressed in vivo to protect against lethal influenza virus infection.a Schematic depicting the generation and transduction of murine tracheal epithelial cell cultures at air–liquid interface (ALI). AAV = adeno-associated virus. b Expression of B3GAT1 in GFP AAV6 and B3GAT1 AAV6 transduced ALI cultures, measured by qRT-PCR (mean with SEM, N = 4 independent cultures). c Representative images of B/Malaysia/2506/2004 (Mal/04) or mock-infected ALI cultures (MOI = 1, 24 HPI). Scale bar is 30 µm. d Quantification of the infected area of ALI cultures from Fig. 4c (mean with SEM, N = 4 independent cultures, *p = 0.0286). e End point titer of Mal/04 infections in ALI cultures (MOI = 1, 24 HPI), measured by plaque assays (mean with SEM, N = 4 independent cultures, *p = 0.0286). PFU = plaque-forming units. f Representative images of PR8 or mock-infected ALI cultures (MOI = 10, 24 HPI). Scale bar is 30 µm. g Quantification of the infected area of ALI cultures from Fig. 4f (mean with SEM, N = 4 independent cultures, *p = 0.0286). h End point titer of PR8 infections in ALI cultures (MOI = 10, 24 HPI), measured by plaque assays (mean with SEM, N = 4 independent cultures, *p = 0.0286). i Timeline for the AAV transduction and influenza virus challenge of C57BL/6J mice. vg = viral genomes. j qRT-PCR for B3GAT1 RNA in mouse lung homogenates (25 days post transduction, mean with SEM, N = 5 mice, *p = 0.0079). k Bodyweight of mice following B3GAT1 AAV6 and GFP AAV6 transduction (N = 5 mice). l Bodyweight (top) and survival (bottom) of B3GAT1 AAV6 and GFP AAV6 transduced mice following Mal/04 infection (50,000 PFU, N = 5 mice). m Bodyweight (top) and survival (bottom) of B3GAT1 AAV6 and GFP AAV6 transduced mice following PR8 infection (100 PFU, N = 4 mice). n Bodyweight (top) and survival (bottom) of B3GAT1 AAV6 and GFP AAV6 transduced mice following A/Hawaii/70/2019 (HI/19) infection (30,000 PFU, N = 5 mice). Dotted line on all bodyweight plots depicts 75% bodyweight humane endpoint. All bodyweight plots show the mean with SEM. Unless otherwise stated, statistical analyses were performed using a two-tailed Mann-Whitney U test. Survival plots p values were determined using the logrank (Mantel–Cox) test. * indicates p < 0.05. Source data are provided as a Source Data file.

Fig 5: Small organic compounds inhibit the binding of the HNK1-specific antibody to the HNK1 mimetic peptide in a concentration-dependent manner. (a) HNK1-specific antibody was substrate-coated on an enzyme-linked immunosorbent assay (ELISA) plate and preincubated with different concentrations (1–200 μM) of compounds from libraries or the negative control compound tacrine. Biotinylated HNK1 mimetic peptide was then added and detected by horseradish peroxidase (HRP)-coupled streptavidin followed by development with o-phenylenediamine dihydrochloride (OPD) substrate to produce a colored product that was quantified in the ELISA reader. The signal with this peptide without compounds was set to 100%. The graph shows average relative inhibition of the signal (duplicate wells carried out in three independent experiments ± SEM). All data points below the asterisks are different in inhibition when compared to the negative control compound tacrine (one-way ANOVA, F(31/64) = 4.122, p < 0.0001; Fisher’s PLSD test, * p < 0.05, ** p < 0.01). (b) Structure of the HNK1 mimetic compounds.