Eukaryotic translation initiation factor 5A (eIF5A) is the only cellular protein that contains the unusual amino acid hypusine [N^ε‐(4‐amino‐2‐hydroxybutyl)lysine]. Vertebrates carry two genes that encode two eIF5A isoforms, eIF5A‐1 and eIF5A‐2, which, in humans, are 84% identical. eIF5A‐1 mRNA (1.3 kb) and protein (18 kDa) are constitutively expressed in human cells. In contrast, expression of eIF5A‐2 mRNA (0.7–5.6 kb) and eIF5A‐2 protein (20 kDa) varies widely. Whereas eIF5A‐2 mRNA was demonstrable in most cells, eIF5A‐2 protein was detectable only in the colorectal and ovarian cancer‐derived cell lines SW‐480 and UACC‐1598, which showed high overexpression of eIF5A‐2 mRNA. Multiple forms of eIF5A‐2 mRNA (5.6, 3.8, 1.6 and 0.7 kb) were identified as the products of one gene with various lengths of 3′‐UTR, resulting from the use of different polyadenylation (AAUAAA) signals. The eIF5A‐1 and eIF5A‐2 precursor proteins were modified comparably in UACC‐1598 cells and both were similarly stable. When eIF5A‐1 and eIF5A‐2 coding sequences were expressed from mammalian vectors in 293T cells, eIF5A‐2 precursor was synthesized at a level comparable to that of eIF5A‐1 precursor, indicating that the elements causing inefficient translation of eIF5A‐2 mRNA reside outside of the open reading frame. On sucrose gradient separation of cytoplasmic RNA, only a small portion of total eIF5A‐2 mRNA was associated with the polysomal fraction, compared with a much larger portion of eIF5A‐1 mRNA in the polysomes. These findings suggest that the failure to detect eIF5A‐2 protein even in eIF5A‐2 mRNA positive cells is, at least in part, due to inefficient translation.