Fig 2: The PAS1 domain is critical for NPAS3 aggregation, as determined by the expression of Flag-tagged NPAS3 fragment in SH-SY5Y cells. (a) Schematic of the fragments of NPAS3 expressed. (b) Expression of the flag-tagged NPAS3 fragments confirmed by Western blotting in HEK293 cells. A mock transfection with no plasmid is used as a negative control. (c–g) Typical localization patterns of these NPAS3 fragments when over-expressed in SH-SY5Y cells: fragments containing amino acids 1-111 (c) or 1-156 (d) (bHLH domain with or without linker region) are seen in the nucleus, while those containing amino acids 1-208 (e,f) or 1-354 (g) (bHLH and PAS1 domains, with or without linker region) are seen in the cytoplasm and in some instances form discernable aggregates. The nucleus (DAPI) and actin cytoskeleton (phalloidin) are shown for context. All amino acid numbers refer to the 933 long amino acid splice variant of human NPAS3 (NP_001158221.1). (h) Quantification of the localization patterns of these aggregates in a blinded assay (5–6 experimental replicates per plasmid, with 4–12 cells being counted in each). n.s.: not significant.

Fig 3: Raised levels of NPAS3 in the serum of a subset of schizophrenia patients. (a) Sample Western blot of whole serum samples from patients with schizophrenia (S) or depression (D), as well as control individuals (C). Stained with an anti-NPAS3 antibody, showing 14 samples (all samples in Supplementary Figure S7). Each lane was loaded with exactly 10 µL of serum. (b) Amount of 130 kDa NPAS3 detected in all serum samples, normalized to the mean over the entire experiment. (c) Amount of 130 kDa NPAS3 detected in only those samples that showed more than mean levels of 130 kDa NPAS3. NPAS3 signal in each category is shown relative to the average NPAS3 signal over all samples. (d) Patient symptom severities, shown as total scores on the positive, negative and general psychopathology scales of PANSS [45], of schizophrenia patients with above background levels NPAS3 in blood serum (n = 7, “NPAS3 positive”) compared to other schizophrenia patients in the study (n = 41, “NPAS3 negative”). There are no significant differences between the two groups. (e) Average number of first- or second-degree relatives of these patients with known diagnosis of mental illness. * p < 0.05, ** p < 0.01, ns: not significant.

Fig 4: Mislocalization of NPAS3 to the cytoplasm does not require the V304I mutation, as determined in SH-SY5Y neuroblastoma cells. (a) One-dimensional schematic of the NPAS3 protein, showing major domains and the location of the V304I mutation. Abbreviations: bHLH: Basic helix-loop-helix domain, PAS: Per-Amt-Sim domain, TAD: C-terminal activator domain. (b) Expression of NPAS3 wild type (WT) and V304I mutant constructs, confirmed by Western blotting in HEK293 cells. Negative controls are mock transfection and transfection with empty vector only, (c–f) Typical localization patterns of NPAS3, both wild type and mutant V304I when over-expressed in SH-SY5Y cells, with each showing a mixture of nuclear and cytoplasmic localization: (c) WT in the nucleus, (d) WT in the nucleus and cytoplasm, (e) V304I in the nucleus, (f) V304I in the cytoplasm. The nucleus (DAPI) and actin cytoskeleton (phalloidin) are shown for context. (g) Quantification of the localization patterns from parts c-f in a blinded experiment. Cells were transfected with either WT or V304I NPAS3 and then exposed to either sodium arsenide (which induces oxidative stress) or an equivalent volume of DMSO as a control. 8 experimental replicates were performed for each combination of plasmid/stress status, with 7–10 cells being counted in each replicate.