AP1S3 mutations are associated with pustular psoriasis and impaired Toll-like receptor 3 trafficking

Abstract

Adaptor protein complex 1 (AP-1) is an evolutionary conserved heterotetramer that promotes vesicular trafficking between the trans-Golgi network and the endosomes. The knockout of most murine AP-1 complex subunits is embryonically lethal, so the identification of human disease-associated alleles has the unique potential to deliver insights into gene function. Here, we report two founder mutations (c.11T>G [p.Phe4Cys] and c.97C>T [p.Arg33Trp]) in AP1S3, the gene encoding AP-1 complex subunit σ1C, in 15 unrelated individuals with a severe autoinflammatory skin disorder known as pustular psoriasis. Because the variants are predicted to destabilize the 3D structure of the AP-1 complex, we generated AP1S3-knockdown cell lines to investigate the consequences of AP-1 deficiency in skin keratinocytes. We found that AP1S3 silencing disrupted the endosomal translocation of the innate pattern-recognition receptor TLR-3 (Toll-like receptor 3) and resulted in a marked inhibition of downstream signaling. These findings identify pustular psoriasis as an autoinflammatory phenotype caused by defects in vesicular trafficking and demonstrate a requirement of AP-1 for Toll-like receptor homeostasis.

Figure 1

Identification of AP1S3 Mutations in Individuals Affected by Pustular Psoriasis (A) Clinical presentation of ACH with active sterile pustules, scaling, nail dystrophy, and digit tapering. (B) Protein sequence alignments demonstrating the evolutionary conservation of the Phe4 and Arg33 residues. (C) Representative chromatograms of the c.11T>G (p.Phe4Cys) and c.97C>T (p.Arg33Trp) mutations. The position of the nucleotide change is highlighted by an asterisk.

Figure 2

Structural Impact of the p.Phe4Cys and p.Arg33Trp Substitutions (A) Comparison between WT (AP-1 σ1C^WT, cyan) and p.Phe4Cys (AP-1 σ1C^p.Phe4Cys, blue) proteins. The residues within 6 Å of AP-1 σ1C^WT Phe4 and AP-1 σ1C^p.Phe4Cys Cys4 are shown in stick representation and are colored in light green and green, respectively. The hydrophobic interactions between Phe4 and its neighboring residues are represented by red dotted lines. (B) Comparison between WT (AP-1 σ1C^WT, cyan) and p.Arg33Trp (AP-1 σ1C^p.Arg33Trp, orange) proteins. AP-1 σ1C residues Arg33 and Trp33, together with AP-1 σ1A residue Glu232 (which lies within 6 Å of Arg33), are shown in stick representation. (C) Immunoblot demonstrating the reduced accumulation of p.Phe4Cys proteins. WT and mutant pcDNA3.1-Myc-AP1S3 constructs were transfected alongside a pcDNA3.1-Myc-RNF114 plasmid (which was used as a transfection-efficiency and loading control) into HEK293 cells. Protein extracts were analyzed by immunoblot with an anti-Myc antibody (MA1-21316, 1:1,000 dilution, Thermo Scientific). The image is representative of the results obtained in two independent transfections. (D) Densitometric analysis of the immunoblots described in (C). The ratio between the intensities of the AP-1 σ1C and RNF114 bands was derived with ImageJ. Data are shown as means ± SD of two independent experiments. ^∗p < 0.05 (unpaired t test).

Figure 3

AP1S3 Deficiency Affects TLR-3 Trafficking and Downstream Signaling (A) Stable AP1S3 knockdown in HaCaT immortalized keratinocytes. Approximately 3 × 10⁵ HaCaT cells were transduced with the supernatants of HEK293 cultures (5 × 10⁵ cells) transfected with 10 μg packaging plasmids (2 μg VSV-G and 8 μg p8.91, Addgene 8454 and 8455) and 2 μg control (ctr; nonsilencing) or silencing shRNA construct. After selection in 10 μg/ml puromycin, the knockdown efficiency was measured by real-time PCR (Table S5) with PPIA expression as an internal control (Life Technologies TaqMan assay 4326316E). Data are presented as mean ± SD of two independent RNA extractions. ^∗∗∗p < 0.001 (unpaired t test). (B) AP1S3 knockdown disrupted trafficking of endogenous TLR-3 in HaCaT keratinocytes. Control and knockdown cells were stimulated with 25 μg/ml poly(I:C) (Invivogen) for 18 hr. The amounts of endosomal (N-ter fragment) and newly synthesized (full-length [FL]) TLR-3 were assessed by immunoblotting with Thermo Fisher Scientific PA5-20184 antibody (1:500 dilution). The intensities of the two bands were measured with ImageJ, and the ratio between the N-ter and FL isoforms was determined for each sample. Bold, underlined font indicates the N-ter/FL ratios observed after poly(I:C) stimulation. β-actin levels determined with Cell Signaling antibody 4970 (1:1,000 dilution) are shown as a loading control. (C) AP1S3 knockdown resulted in reduced TLR-3 signaling. Control and knockdown cells were stimulated with 25 μg/ml poly(I:C) for 3 hr, and IFNB1 induction was measured by real-time PCR (Life Technologies TaqMan assay 4331182) with PPIA expression as an internal control. The data are representative of the results obtained in at least two independent experiments and are shown as mean ± SD of two technical duplicates. ^∗∗p < 0.01 (unpaired t test). (D) Stable AP1S3 knockdown in HEK293 cells. Approximately 3 × 10⁵ cells were transduced with lentiviral shRNA, as described above. The knockdown efficiency was measured by real-time PCR with PPIA expression as an internal control. Data are presented as mean ± SD of technical triplicates. ^∗∗p < 0.001 (unpaired t test). The following abbreviation is used: ctr, control cells transduced with a nonsilencing shRNA. (E) Compared to AP1S3-knockdown cells overexpressing WT AP-1 σ1C, AP1S3-knockdown cells overexpressing p.Arg33Trp protein showed reduced TLR-3 processing. Left panel: control and knockdown HEK293 cells were transfected with TLR3 cDNA (Addgene 13084) for 48 hr. The amounts of endosomal (N-ter) and newly synthesized (FL) receptor were assessed by immunoblotting, and the ratio between the two isoforms was determined by densitometry. Right panel: knockdown HEK293 cells were transfected with WT or mutant AP1S3 cDNA constructs (encoding Arg33 or Trp33, respectively) for 24 hr and with TLR3 cDNA for a further 48 hr. The amounts of N-ter and FL receptor were monitored by immunoblotting and densitometry. β-actin levels are shown as a loading control. The images are representative of the results obtained from two independent experiments.