BACE2 distribution in major brain cell types and identification of novel substrates

mRNA expression is assigned as very low (+), low (++), medium (+++), high (++++), and very high (+++++). Polr2a, Ppib, and Ubc are housekeeping brain-specific genes previously studied with RNAscope, and Syp and Glast are known cell type markers. The expression levels are also reported from Brain RNA Seq database (http://web.stanford.edu/group/barres_lab/brain_rnaseq.html). FPKM, fragments per kilobase of transcript per million mapped reads.

Volcano plot of proteomic analysis of CSF from Bace1^−/− and dKO mice (n = 5). For each relatively quantified protein (representing a dot on the plot), the log2-transformed LFQ intensity ratios of dKO/Bace1^−/− was plotted against the −log10-transformed t test P-values. Proteins with a t test P-value <0.05 are marked with open red circles, whereas proteins with P-value >0.05 are shown in blue. Labeled in bold black letters are the four previously identified BACE1 substrate candidates. In green are other transmembrane proteins, not previously identified as substrates of BACE1 or BACE2. No corrections for multiple hypothesis testing were applied in this discovery experiment.

(A) Volcano plot of proteomic analysis of conditioned medium from Bace1^−/− cultured glia treated with BACE inhibitor CpJ or vehicle (n = 3). For each relatively quantified protein (representing a dot on the plot), the −log10-transformed t test P-value was plotted against the log2-transformed LFQ intensity ratios of CpJ/vehicle. Proteins with a t test P-value <0.05 are marked with open red circles, whereas proteins with P-value >0.05 are shown in blue. Closed red dots labeled with bold letters denote the four substrate candidates that are reduced by more than 30% in the inhibitor-treated samples. No corrections for multiple hypothesis testing were applied in this discovery experiment. (B) Uniprot subcellular locations of the identified proteins. Glycoproteins were defined according to UniProt Keywords. (C) Topology of membrane proteins according to Uniprot subcellular locations. (D) Uniprot Topology of DNER, FGFR1, PLXDC2, and VCAM1 with mapped identified peptides.

(A) VCAM1 in medium and cell lysates of Bace1^−/− glia treated with vehicle or CpJ (10 μM) for 24 h (n = 4). (B) Quantification of VCAM1 shedding into conditioned medium of Bace1^−/− glia treated with vehicle or CpJ for 24 h (paired t test, P = 0.01, n = 4). (C) Quantification of VCAM1 accumulation in cell lysates of Bace1^−/− glia treated with vehicle or CpJ (10 μM) for 24 h (paired t test, P = 0.08, n = 4). (D) VCAM1 shedding in WT, Bace1^−/−, and Bace2^−/− glia (n = 3). (E) VCAM1 in medium and cell lysates of WT glia treated with vehicle or CpJ (10 μM) for 24 h (n = 3). (F) VCAM1 in medium and cell lysates of Bace2^−/− glia treated with vehicle or CpJ (10 μM) for 24 h (n = 3). (G) DNER in medium and cell lysates of Bace1^−/− glia treated with vehicle or CpJ (10 μM) for 24 h (n = 3). (H) Quantification of DNER shedding into conditioned medium of Bace1^−/− glia treated with vehicle or CpJ for 24 h (paired t test, P = 0.03, n = 3). (I) Quantification of DNER accumulation in cell lysates of Bace1^−/− glia treated with vehicle or CpJ for 24 h (paired t test, P = 0.3, n = 3). (J) DNER shedding in WT, Bace1^−/−, and Bace2^−/− glia (n = 3). (K) DNER in medium and cell lysates of WT glia treated with vehicle or CpJ (10 μM) for 24 h (n = 3). (L) DNER in medium and cell lysates of Bace2^−/− glia treated with vehicle or CpJ (10 μM) for 24 h (n = 3).

(A) FGFR1 in lysates and medium of COS-1 cells overexpressing FGFR1 alone (lane 2), FGFR1 with BACE1 (lanes 3 and 4), and FGFR1 with BACE2 (lanes 5 and 6). Lanes 4 and 6 were treated with inhibitor CpJ (10 μM). Lane 1 was mock-transfected with an empty vector. Representative for three experiments. (B) PLXDC2 in medium (top panel) and lysates (second panel from the top) of COS-1 cells overexpressing PLXDC2 alone (lane 2), PLXDC2 with BACE1 (lanes 3 and 4), and PLXDC2 with BACE2 (lanes 5 and 6). Lanes 4 and 6 were treated with inhibitor CpJ (10 μM). Lane 1 was mock-transfected with an empty vector. Control blots for BACE1, BACE2, and actin are shown for each transfection. Representative for three experiments.

(A) Shed VCAM1 and DNER in triplicates of CSF of 11-mo-old WT, Bace1^−/−, Bace2^−/−, and dKO male mice, as compared with a known BACE1 substrate—SEZ6. (B) SEZ6 is significantly reduced in Bace1^−/− and dKO, but not Bace2^−/− CSF compared with control WT CSF (one-way ANOVA: P = 0.006, P = 0.008, and P = 1.00, respectively; n = 3). (C) One-way ANOVA revealed no significant differences between shed DNER in Bace1^−/−, Bace2^−/−, and dKO CSF, as compared with WT (P = 1.00, P = 1.00, and P = 0.77, respectively; n = 3). (D) No significant differences were seen between shed VCAM1 in Bace1^−/−, Bace2^−/−, dKO, and WT CSF by one-way ANOVA (P = 0.76, P = 1.00, and P = 1.00, respectively; n = 3).

(A) DNER in the soluble TBS fraction and total cell lysates of homogenized cortex from 11-mo-old WT, Bace1^−/−, Bace2^−/−, and dKO male mice. Actin is used as a loading control. (B) Quantification of soluble DNER in the TBS fraction showed no significant differences between the four genotypes as analyzed by one-way ANOVA (P = 1.00 for Bace1^−/− versus WT, P = 1.00 for Bace2^−/− versus WT, and P = 1.00 for dKO versus WT). (C) VCAM1 in the soluble TBS fraction and total cell lysates of homogenized cortex from 11-mo-old WT, Bace1^−/−, Bace2^−/−, and dKO male mice. Actin is used as a loading control. (D) Quantification of soluble VCAM1 in the TBS fraction showed no significant differences between the four genotypes as analyzed by one-way ANOVA (P = 0.31 for Bace1^−/− versus WT, P = 0.59 for Bace2^−/− versus WT, and P = 1.00 for dKO versus WT). (E, F) No significant differences were observed in full-length levels of DNER and VCAM1 in total cell lysates (one-way ANOVA of FL DNER: P = 0.19 for Bace1^−/− versus WT, P = 0.25 for Bace2^−/− versus WT, and P = 0.25 for dKO versus WT; one-way ANOVA of FL VCAM1: P = 0.16 for Bace1^−/− versus WT, P = 1.00 for Bace2^−/− versus WT, and P = 1.00 for dKO versus WT). (G) VCAM1 in the soluble TBS fraction and total cell lysates of homogenized subdissected ventral hippocampi from 4-mo-old WT, Bace2^−/−, and dKO male mice. Actin is used as a loading control. (H) VCAM1 in the soluble TBS fraction and total cell lysates of homogenized subdissected subventricular zone from P16 WT, Bace1^−/−, Bace2^−/−, and dKO male mice. Actin is used as a loading control. VCAM1 (I) and DNER (J) in the soluble TBS fraction and membrane fractions of homogenized brain hemispheres from 4-mo-old Bace1^−/− male mice treated with CpJ or vehicle. PSEN1 membrane protein was used as a loading control.

(A) VCAM1 in medium and lysates of primary mixed glia culture treated with murine recombinant TNF (10 ng/ml) or IL-1β (10 ng/ml) for 8 h and 24 h (n = 3). Control blots for BACE1, BACE2, and actin are shown. (B) VCAM1 in medium and lysates of primary mixed glia cells treated with murine recombinant TNF (10 ng/ml) for 24 h, or treated with vehicle or CpJ inhibitor (10 μM) (n = 3). Control blots for GFAP and actin are shown. (C) VCAM1 in triplicates of CSF, TBS fraction, and total cell lysates of cortices from 11-mo-old WT male controls injected with PBS or treated with 250 μg/kg TNF. Actin is used as a loading control. (D) VCAM1 shedding into CSF is up-regulated upon TNF treatment (unpaired t test, P = 0.04; n = 3), whereas no changes are observed in full-length VCAM1 or soluble VCAM1 in TBS fraction shown in (C). (E) VCAM1 in triplicates of CSF of 11-mo-old WT or Bace2^−/− mice treated with saline or 250 μg/kg TNF (top panel) and VCAM1 in triplicates of CSF of 11-mo-old Bace1^−/− or Bace2^−/− mice treated with saline or 250 μg/kg TNF (bottom panel). (F) Two-way ANOVA reveal significant differences in shed VCAM1 between treated and untreated WT and Bace1^−/− mice (P = 0.0004 and P = 0.005, respectively; n = 3), but no significant differences in shed VCAM1 between treated and untreated Bace^−/− mice (P = 0.69).

DNER in medium and lysates of primary mixed glia culture treated with murine recombinant TNF (10 ng/ml) or IL-1β (10 ng/ml) for 8 h and 24 h (n = 3). Control blots for BACE2 and actin are shown.