In order to quantitate the pathways by which liver glycogen is repleted, we administered [1-13C]glucose by gavage into awake 24-h fasted rats and examined the labeling pattern of 13C in hepatic glycogen. Two doses of [1-13C]glucose, 1 and 6 mg/g body wt, were given to examine whether differences in the plasma glucose concentration altered the metabolic pathways via which liver glycogen was replenished. After 1 and 3 h (high-dose group) and after 1 and 2 h (low-dose group), the animals were anesthetized and the liver was quickly freeze-clamped. Liver glycogen was extracted and the purified glycogen hydrolyzed to glucose with amyloglucosidase. The distribution of the 13C-label was subsequently determined by 13C-nuclear magnetic resonance spectroscopy. The percent 13C enrichment of the glucosyl units in glycogen was: 15.1 +/- 0.8%(C-1), 1.5 +/- 0.1%(C-2), 1.2 +/- 0.1%(C-3), 1.1 +/- 0.1%(C-4), 1.6 +/- 0.1%(C-5), and 2.2 +/- 0.1%(C-6) for the high-dose study (n = 4, at 3 h); 16.5 +/- 0.5%(C-1), 2.0 +/- 0.1%(C-2), 1.3 +/- 0.1%(C-3), 1.1 +/- 0.1%(C-4), 2.2 +/- 0.1%(C-5), and 2.4 +/- 0.1%(C-6) in the low-dose study (n = 4, at 2 h). The average 13C-enrichment of C-1 glucose in the portal vein was found to be 43 +/- 1 and 40 +/- 2% in the high- and low-dose groups, respectively. Therefore, the amount of glycogen that was synthesized from the direct pathway (i.e., glucose----glucose-6-phosphate----glucose-1-phosphate----UDP-glucose---- glycogen) was calculated to be 31 and 36% in the high- and low-dose groups, respectively. The 13C-enrichments of portal vein lactate and alanine were 14 and 14%, respectively, in the high-dose group and 11 and 8%, respectively, in the low-dose group. From these enrichments, the minimum contribution of these gluconeogenic precursors to glycogen repletion can be calculated to be 7 and 20% in the high- and low-dose groups, respectively. The maximum contribution of glucose recycling at the triose isomerase step to glycogen synthesis (i.e., glucose----triose-phosphates----glycogen) was estimated to be 3 and 1% in the high- and low-dose groups, respectively. In conclusion, our results demonstrate that (a) only one-third of liver glycogen repletion occurs via the direct conversion of glucose to glycogen, and that (b) only a very small amount of glycogen synthesis can be accounted for by the conversion of glucose to triose phosphates and back to glycogen; this suggests that futile cycling between fructose-6-phosphate and fructose-1,6-diphosphate under these conditions is minimal. Our results also show that (c) alanine and lactate account for a minimum of between 7 and 20% of the glycogen synthesized, and that (d) the three pathways through which the labeled flux is measured account for a total of only 50% of the total glycogen synthesized. These results suggest that either there is a sizeable amount of glycogen synthesis via pathway(s) that were not examined in the present experiment or that there is a much greater dilution of labeled alanine/lactate in the oxaloacetate pool than previously appreciated, or some combination of these two explanations.