Blood-borne nutrient supply is critical to fuel muscle contraction and limits exercise performance. A silencer region (I-allele) within intron 16 of the gene for the major regulatory enzyme of vasoconstriction, angiotensin converting enzyme (ACE), is implicated in phenotypic variation of human performance and its trainability. The muscular mechanisms contributing to ACE-modulated exercise performance is not known. We hypothesized that modified substrate supply and aerobic metabolism in skeletal muscle underlies the modulation of maximal aerobic exercise performance by the genetic silencing of ACE expression. This was exposed using a genetical-metabolomics approach. Human subjects, who gave informed consent, were recruited to characterize ACE-dependence in muscle makeup and muscle’s metabolic response to exercise. Aerobic performance was assessed by ergospirometry. Muscle biopsies were collected (under lidocaine as local anaesthetic), from the major extensor muscle vastus lateralis, before and 30 min after a bout of maximal aerobic exercise at a standardized intensity on an ergometer (Jaeger). Biopsies of untrained healthy subjects (n=20) were subjected to transcript profiling and ultra-structural analysis, carried out as described (Schmutz et al 2010). Muscle metabolites of untrained (n=10) and trained (n=10) subjects were extracted from cryo-sectioned samples into a polar (methanol) and non-polar (chloroform) phase and assessed by liquid chromatography – mass spectrometry (LC-MS) and direct infusion mass spectrometry (DIMS), respectively, as described in (Dunn et al 2008; Brown et al 2009). ACE-DD genotypes demonstrated elevated muscle capillarity (+17.8%; p<0.05) and increased expression (assessed by significance analysis of microarray – SAM) of transcripts related to vascular remodelling and lipid metabolism compared to ACE-II/ACE-ID genotypes (q-value = 2.39; fold change 1.01 – 2.08). Oxidation of lipid substrates (as assessed by respiration exchange ratio; -30%; p=0.30, ANOVA) and levels of non-polar, but not polar, metabolites were reduced with exercise in the ACE-DD genotype lacking the I-allele (-33%, p=0.06, repeated ANOVA). The selective depletion of the non-polar metabolite class was preserved in trained subjects (-34%, p=0.03). Serum LDL was selectively reduced (-14%, p=0.03) in the genotypes with the I-allele. The observations demonstrate a clear difference in whole-body substrate utilisation between the investigated ACE-genotypes during maximal endurance exercise, which relates to muscle metabolite levels. A mismatch between vascular delivery and myocellular turnover of non-polar metabolites upon exercise is suggested to underlie the different trainability of genotypes for the major checkpoint of vascular perfusion, ACE (Montgomery et al 1998).