Background: Spinocerebellar ataxia type 2 (SCA2) is an autosomal dominant disorder with progressive degeneration of cerebellar Purkinje cells and selective loss of neurons in the brainstem. This neurodegenerative disorder is caused by the expansion of a polyglutamine domain in ataxin-2. Ataxin-2 is composed of 1312 amino acids, has a predicted molecular weight of 150-kDa and is widely expressed in neuronal and non-neuronal tissues. To date, the putative functions of ataxin-2 on mRNA translation and endocytosis remain ill-defined. Differential splicing with a lack of exons 10 and 21 was described in humans, and additional splicing of exon 11 in mice. In this study, we observed that the molecular size of transfected full-length wild-type ataxin-2 (22 glutamines) is different from endogenous ataxin-2 and that this variation could not be explained by the previously published splice variants alone.Methods: Quantitative immunoblots and qualitative reverse-transcriptase polymerase-chain-reaction (RT-PCR) were used to characterize isoform variants, before sequencing was employed for validation. Results: We report the characterization of further splice variants of ataxin-2 in different human cell lines and in mouse and human brain. Using RT-PCR from cell lines HeLa, HEK293 and COS-7 throughout the open reading frame of ataxin-2 together with PCR-sequencing, we found novel splice variants lacking exon 12 and exon 24. These findings were corroborated in murine and human brain. The splice variants were also found in human skin fibroblasts from SCA2 patients and controls, indicating that the polyglutamine expansion does not abolish the splicing. Conclusions: Given that Ataxin-2 interacts with crucial splice modulators such as TDP-43 and modulates the risk of Amyotrophic Lateral Sclerosis, its own splice isoforms may become relevant in brain tissue to monitor the RNA processing during disease progression and neuroprotective therapy.