The choline analog, N-amino-N,N-dimethylaminoethanol (NADe), was fed ad libitum (chloride salt; 0.5%) to weanling rats in a low choline, low methionine synthetic diet. Control rats were fed choline chloride (0.5%) in place of NADe. Initial observation and behavioral screen tests of grasp strength, startle reflex, righting reflex, analgesia (hot plate test) and body temperature did not reveal any toxic effects caused by NADe, although both experimental and control groups gained weight more slowly than rats fed standard lab chow. After 25 days on the diet, the performance of rats fed NADe in a one-trial passive avoidance test was significantly impaired compared to control rats. There was no difference between experimental and control rats in sensitivity to foot shock or in activity monitored in a closed field. A subjective, 6-component behavioral rating scale indicated rats fed NADe were resistant to handling but not aggressive. These behavioral results were similar in two separate feeding experiments using deuterium-labeled and unlabeled NADe. The twitch response of isolated rat phrenic nerve-diaphragms during stimulation did not show any impairment of neuromuscular function in rats fed NADe. Receptor binding experiments indicated there were no differences between experimental and control rats in tritiated quinuclidinyl benzilate [( 3H]QNB) binding capacity in cortex, heart and ileum. Competitive [3H]QNB binding with carbachol indicated there was no difference in the IC50's measured in cortex homogenates. Acetylcholinesterase (AChE) and choline acetyltransferase (ChAT) activities in cortex were similar in experimental and control groups.(ABSTRACT TRUNCATED AT 250 WORDS)