Splicing of human GABAB receptor subunit 1 (GABAB1) in non-alcoholic and alcoholic brains

Gamma-aminobutyric acid type B (GABAB) receptor is a G protein coupled receptor (GPCR) that mediates decreased neural activity. It has two subunits, GABAB1 and GABAB2. Previous complementary DNA (cDNA) microarray data showed strong GABAB1 signals from human prefrontal cortex using an intron 4 region probe, and these studies indicated that novel intron 4 containing GABAB1 splicing variants exist. We cloned GABAB1k, l, m, and n including mouse GABAB1j. Expression of these variants are much lower than other major known splicing variants, but GABAB1k, l, m, and n levels are similar across brain tissues. GABAB1l and GABAB1m impair GABAB receptor induced function. To better define GABAB1 splicing in alcoholic brains, whole transcriptome shortgun sequencing (RNA-seq) experiments were proposed. Due to the complexity of GABAB1 splicing, we used gene specific libraries as well as whole transcriptome libraries to maximize GABAB1 specific splicing junction search. The splicing junction search data found that GABAB1 gene is 2 to 3 times longer than the previous known gene length. Extremely low expression at 5’ end exons was found, and GABAB1 exons were grouped based on expression levels. Chronic alcohol altered exon/intron expression and splicing junctions more than overall gene expression. Decreased exon expression at a GABA binding site, a transmembrane domain (TM), and a microRNA (miRNA) binding site may diminish the normal GABAB1 transcript population and compromise signal transduction following chronic alcohol exposure. This may explain why GABAB receptor agonists have therapeutic benefit in treating alcoholism. During the sequence mapping, read pile-ups and gaps were found from whole transcriptome libraries in known exons. These may prevent single nucleotide polymorphism (SNP) and splicing junction identification and gene expression calculations. Sequence analysis found sequence biases from their mapped reads. The major sequence biases were from RNaseIII RNA fragmentation and T4 polynucleotide kinase (T4PNK) reaction. Heat fragmentation and OptiKinase treatment removed the read pile-ups and gaps including the sequence biases. The identification of RNaseIII target sequences can be incorporated into methods of miRNA gene prediction. These data showed the complexity of GABAB1 receptor splicing and the perturbation of splicing by chronic alcohol abuse demonstrate the power of RNA-seq to provide new insight into gene expression and the role of GABAB receptors in alcoholism. In addition, many other important brain genes may have unexplored splicing variants which will be important for alcoholism and other psychiatric diseases. Also, new RNA-seq library constructions improved the quality of gene expression studies.