The whey acidic protein (WAP) is a major milk protein in mice, rats, and rabbits but has not been found in milk of livestock including swine. To determine whether mammary gland regulatory elements from the WAP gene function across species boundaries and whether it is possible to qualitatively alter milk protein composition, we introduced the mouse WAP gene into the genome of swine. Three lines of transgenic swine were analyzed, and mouse WAP was detected in milk from all lactating females at concentrations of about 1 g/liter; these levels are similar to those found in mouse milk.Expression of the corresponding RNA was specific to the mammary gland. Our results suggest that the molecular basis of mammary-specific gene expression is conserved between swine and mouse. In addition the WAP gene must share, with other milk protein genes, elements that target gene expression to the mammary gland. Mouse WAP accounted for about 3% of the total milk proteins in transgenic pigs, thus demonstrating that it is possible to produce high levels of a foreign protein in milk of farm animals.Milk protein genes are transcribed in the mammary gland of lactating animals, and the encoded proteins are secreted in large quantities into milk. The whey acidic protein (WAP) is an abundant milk protein in mice (1, 2) but has not been found in swine or other livestock. Expression of the WAP gene is confined to the mammary gland (2, 3) and is under the control of steroid and peptide hormones as well as other developmental signals during pregnancy (4-6).By targeting synthesis of foreign proteins to the mammary gland of transgenic animals, it should be possible to produce valuable proteins on a large scale in milk (7,8 Swine were chosen for these studies because they offer both economy in animal resources and time when compared to ruminantia as a transgenic animal model and because the questions being addressed did not require harvesting large quantities of milk that would be more easily obtained from dairy animals such as cows, goats, or sheep. The two primary constraints in any large animal transgenic project are the number of fertilized ova obtainable and the number of embryo recipients available. On average it is possible to recover 2-3 times more injectable ova per donor gilt than can be collected from a cow, doe, or ewe. The efficiency of producing expressing transgenic pigs or sheep per injected ovum is about 0.3% (calculated from refs. 11 and 12). Though a live-born-expressing transgenic calf has not been reported, a larger number of ova will probably be required to produce an expressing transgenic cow (13). Furthermore, because swine are polytocous, a recipient sow can carry 5 times as many fetuses as a cow, doe, or ewe. Additionally, the generation interval of swine is -11 months, whereas that of goats is between 11 and 21 months and that of cattle at least 24 months. Considering all of these factors, the use of swine rather than cows, goats, or sheep requires one-sixth the number of animals, with results obtainable in less th...