The dietary selenium recommendation for turkeys of 0.2 microg Se/g is higher than for many other species. Liver glutathione peroxidase-1 (Gpx1) activity levels determined using hydrogen peroxide (H(2)O(2)) in previous studies suggest that 0.2 microg Se/g may still be too low and that some of this Gpx1 activity might be due to phospholipid hydroperoxide Gpx (Gpx4). Thus we separated Gpx1 from Gpx4 by chromatography, demonstrated that 47% of the H(2)O(2) activity in Se-adequate turkey liver was due to Gpx4, and determined a factor for calculation of each activity. Day-old male poults were fed an Se-deficient torula diet (0.007 microg Se/g) supplemented with graded levels of Se (0-0.5 microg Se/g) for 27 days. Final body weights indicated a minimum Se requirement for growth of 0.05 microg Se/g. The liver had the highest Gpx4 activity in Se-adequate poults, and Gpx4 activity in Se-deficient liver decreased to 5% of Se-adequate levels, with an Se requirement of 0.29 microg Se/g. Liver Gpx1, gizzard Gpx1 and gizzard Gpx4 activities also had Se requirements of 0.28-0.30 microg Se/g, collectively yielding an Se requirement of 0.3 microg Se/g, which is three times higher than the requirements found in comparable rodent studies. We also sequenced partial cDNA clones for turkey Gpx1 (GQ502186) and Gpx4 (GQ502187), and found >60% identity with rodents and humans and >90% identity with chickens. Ribonuclease protection analysis showed that Gpx4 mRNA levels decrease substantially in Se-deficient turkey liver, unlike in rodents. These underlying differences in selenoprotein molecular biology may explain the elevated dietary Se requirements of turkeys.