In non-lymphoid cells such as HeLa cells, ectopic expression of the lymphocyte-specific transcription factor Oct-2A can activate reporter genes whose promoters consist of a single octamer sequence (ATTTGCAT) upstream of a TATA box. While the factor is strongly active in a promoter position, it tails as an enhancer factor: an enhancer consisting of multiple copies of the octamer sequence placed downstream of the reporter gene is not active in HeLa cells, even at high concentration of Oct-2A. In B lymphoid cells, however, the same enhancer is highly active. This could mean that an additional factor is required for enhancer activation in B cells. Furthermore, we have tested the transcriptional activation potential of Oct-2A with a series of N-terminal and C-terminal deletions. We show that a glutamine-rich domain near the N-terminus is required for full activity. Otherwise, large segments of the N-terminal half or the entire C-terminal region are dispensable in our assay, as long as the deletions do not impinge on the conserved POU domain which is sufficient for DNA binding. While N-terminal and C-terminal regions can functionally compensate for each other, a combined deletion that only retains the POU domain is a strong down mutation. We also find that activity depends on the promoter structure of the reporter gene: the POU domain by itself shows some activity with a promoter where the octamer sequence is located very close to the TATA box, but no activity with another promoter construction where the octamer sequence is located further upstream. The two promoters also respond differently to the deletion of the glutamine-rich stretch important for transcriptional activation. From these experiments we consider it likely that the natural octamer factor variants can selectively activate the different naturally occurring octamer-containing promoters.