Genomic organization and expression analysis of BACE1 and BACE1-AS. (a) Genomic sequences of BACE1 and BACE1-AS; arrows show the direction of transcription. BACE1 exons are depicted as vertical bars and marked E1–E9. Human BACE1-AS is transcribed from the same region in chromosome 11, but on the opposite strand. Yellow highlighted exons are the overlapping region (104 base pairs) of BACE1 and BACE1-AS, which are conserved across species. Sites numbered 1, 3 and 4 are BACE1 siRNAs target sites, and site 2 is the northern blot probe site. Sites 5, 6 and 7 are the target sites of the BACE1-AS siRNAs and site 8 is the RT-PCR probe target region, which are all in the non-overlapping part of the BACE1-AS transcript. Sites 5 and 9 represent the primers for 3′ and 5′ RACE, respectively. (b,c) RACE sequencing data revealed that BACE1-AS contains cap structure and a poly-A tail and that this transcript undergoes differential splicing in both human and mouse. The yellow highlighted segments are the overlap region to the BACE1 sense transcript and the green highlighted segments are additional nucleotides observed from our sequencing data. Point mismatches to the genomic sequence are indicated by stars (*) for A to G and crosses (†) for C to T changes. Nt, nucleotides. (d) Expression of BACE1-AS, BACE1 and APP mRNA in human cortical neurons (HCN1A) compared to glial cells (M059K). (e) Northern blot expression analysis of BACE1 (top) and BACE1-AS (bottom) in ten human tissues. S-muscle, skeletal muscle

Bace1-AS regulates Bace1 in vivo. (a–e) Synthetic unmodified siRNAs designed to target the nonoverlapping region of either Bace1 or Bace1-AS and a control siRNA were constantly infused into three groups of mice over a period of two weeks. The siRNAs directed against either Bace1 or Bace1-AS but not the control siRNA resulted in decrease in both Bace1 and Bace1-AS levels (P < 0.0001) in cortex (a), striatum (b), dorsal hippocampus (c) and ventral hippocampus (d). In the cerebellum (e) both transcripts were unchanged, as expected for a tissue that is not directly connected to the third ventricle of the brain. (f) Western blot showing decreases of Bace1 protein abundance in the ventral hippocampus but not in the cerebellum after in vivo treatment with siRNA against either Bace1 or Bace1-AS.

BACE1-AS increases the stability of BACE1 mRNA. (a) RPA performed on RNA samples from SH-SY5Y cells. Depicted here are RT-PCR results from two sets of primers and probes covering overlapping and nonoverlapping regions of BACE1 mRNA. The overlapping region of BACE1 transcript is protected from degradation by RNase A+T, suggesting RNA duplex formation. (b) Stability of BACE1 and BACE1-AS transcripts over time was measured by RT-PCR relative to time 0 after blocking new RNA synthesis with α-amanitin (50 μM) in HEK293T cells. BACE1-AS showed a shorter half-life than BACE1 and ACTB. 18s RNA, which is a product of RNA polymerase I, was unchanged. (c) The stability of BACE1 mRNA was measured in stably transfected HEK293T cells expressing a BACE1-AS shRNA and a second cell line expressing a negative control shRNA. The stability of BACE1 mRNA was decreased in cells expressing BACE1-AS shRNA relative to the control cell line (P < 0.01). (d) The stability of BACE1 mRNA increased in HEK293T cells overexpressing BACE1-AS in comparison to a cell line transfected with an empty vector.

BACE1-AS regulates BACE1 mRNA and protein expression in vitro. (a) Targeting BACE1-AS transcript with three different siRNAs caused decreases (P < 0.0001) in both BACE1 and BACE1-AS transcripts in neuroblastoma cells (SH-SY5Y). (b) Stable transfection of HEK293T cells with shRNA for BACE1, BACE1-AS and a control shRNA shows that knockdown of BACE1 for an extended period of time leads to reduction of BACE1-AS and knockdown of BACE1-AS also leads to reduction of BACE1 mRNA levels (P < 0.001). (c) HEK-SW cells were transfected with 100 pM, 500 pM, 5 nM, 10 nM or 20 nM BACE1-AS siRNA. BACE1-AS knockdown ranged from 10–60%, and BACE1 downregulation ranged between 10–50% with increasing concentrations of siRNA. (d,e) Western blot showing that knockdown of either BACE1 or BACE1-AS with siRNA or shRNA leads to reduction of the BACE1 protein. (f,g) Overexpression of BACE1-AS but not an empty control vector leads to increased BACE1 mRNA (P < 0.001) and protein concentrations. (h) HEK-SW cells were transfected with siRNA for BACE1, BACE1-AS or a control siRNA and analyzed for Aβ 1–40, Aβ 1–42, sAPPα and total APP concentrations by ELISA. Aβ 1–40 and Aβ 1–42 abundance was reduced (P < 0.0001) after transfection of siRNA targeting of either BACE1 or BACE1-AS. Total APP or sAPPα, an enzymatic product of α-secretase, were not changed.

Effect of cell stressors on BACE1 and BACE1-AS. (a) HEK-SW cells were exposed to cell stressors. BACE1-AS and BACE1 transcripts were measured by RT-PCR and normalized to ACTB as an endogenous control. Hyperthermia, serum starvation, Aβ 1–42, hydrogen peroxide or glucose shock caused elevation of BACE1-AS and, to a lesser degree, elevation of BACE1 transcripts. Staurosporine or Aβ 1–40 did not increase either transcript level. (b) The 7PA2-CHO cells were previously shown to overproduce Aβ 1–42 dimers and oligomers. Conditioned media from these cells or control parental CHO cells were collected and added to SH-SY5Y cells for 2 h after removal of the regular media. Conditioned media from 7PA2-CHO cells, but not control media, caused the BACE1-AS transcript to relocate to the cytoplasm (P < 0.0001). (c) Exposure of SH-SY5Y cells to 1 μM Aβ 1–42 for 2 h caused an increase in cytoplasmic BACE1-AS (P < 0.001). The nuclear-cytoplasmic ratio was recovered upon removal of the peptides and maintenance of the cells in regular media for 1 h. (d) Exposure to 1 μM Aβ 1–42 peptide caused an elevation in BACE1 protein abundance (P < 0.001). Protein amounts were measured with a β-galactosidase ECA.

BACE1-AS and BACE1 expression is elevated in the brain of individuals with Alzheimer’s disease. (a) The relative quantity of BACE1-AS transcript was elevated by two to three times in parietal cortex and cerebellum of five human subjects with Alzheimer’s disease (AD subjects) (P < 0.0001) as compared to matched control individuals (20 RNA samples). To a lesser degree, BACE1 mRNA was also increased, by approximately 30%. (b) The relative quantity of BACE1-AS transcript was elevated by almost twofold (P < 0.0001) and BACE1 mRNA elevated about 30% in four brain regions (cerebellum, superior frontal gyrus, entorhinal cortex and hippocampus) of 35 individuals with Alzheimer’s disease compared to the average of 35 control individuals (128 RNA samples in total). (c) Scatter plot of BACE1-AS transcript expression in superior frontal gyrus of 17 control subjects and 16 subjects with Alzheimer’s disease. Upregulation (P < 0.0001) of BACE1-AS was observed in subjects with Alzheimer’s disease. (d) Scatter plot of BACE1-AS transcript expression in hippocampus of 11 controls and 13 individuals with Alzheimer’s disease. Upregulation (P < 0.001) of BACE1-AS was observed in the subjects with Alzheimer’s disease. (e,f) BACE1-AS to ACTB ratio (e) and BACE1-AS to BACE1 ratio (f) in hippocampus (Hipp), superior frontal gyrus (SFG), entorhinal cortex (Ectx) of subjects with Alzheimer’s disease compared to control individuals. Both ratios are higher (P < 0.001) in subjects with Alzheimer’s disease. (g) The concentrations of human Aβ 1–42 peptide are elevated in the brains of APP-tg19599 mice (P < 0.0001). (h) Bace1-AS is elevated by 50% (P < 0.0001) in whole brains of APP-tg19599 (APP-tg) mice.