ABSTRACT: Fatty acids produced by H2-metabolizing bacteria are sometimes observed to be more D-depleted than those of photoautotrophic organisms, a trait that has been suggested as diagnostic for chemoautotrophic bacteria. The biochemical reasons for such a depletion are not known, but are often assumed to involve the strong D-depletion of H2. Here, we cultivated the bacterium Cupriavidus necator H16 (formerly Ralstonia eutropha H16) under aerobic, H2-consuming, chemoautotrophic conditions and measured the isotopic compositions of its fatty acids. In parallel with the wild type, two mutants of this strain, each lacking one of two key hydrogenase enzymes, were also grown and measured. In all three strains, fractionations between fatty acids and water ranged from -173‰ to -235‰, and averaged -217‰, -196‰, and -226‰, respectively, for the wild type, SH- mutant, and MBH- mutant. There was a modest increase in ?D as a result of loss of the soluble hydrogenase enzyme. Fractionation curves for all three strains were constructed by growing parallel cultures in waters with ?Dwater values of approximately -25‰, 520‰, and 1100‰. These curves indicate that at least 90% of the hydrogen in fatty acids is derived from water, not H2. Published details of the biochemistry of the soluble and membrane-bound hydrogenases confirm that these enzymes transfer electrons rather than intact hydride (H-) ions, providing no direct mechanism to connect the isotopic composition of H2 to that of lipids. Multiple lines of evidence thus agree that in this organism, and presumably others like it, environmental H2 plays little or no direct role in controlling lipid ?D values. The observed fractionations must instead result from isotope effects in the reduction of NAD(P)H by reductases with flavin prosthetic groups, which transfer two electrons and acquire H+ (or D+) from solution. Parallels to NADPH reduction in photosynthesis may explain why D/H fractionations in C. necator are nearly identical to those in many photoautotrophic algae and bacteria. We conclude that strong D-depletion is not a diagnostic feature of chemoautotrophy.