Inferences on soil biogeochemical processes based on metagenomic profiles is a challenging task due to enormous diversity of soil microbes and the fragile linkage between gene abundance and functioning. Here we used the biological sink of H-2 as a case study to test the hypothesis that [NiFe]-hydrogenase gene distribution and expression profiles explain variations in H-2 oxidation rate measured in soil collected in poplar monoculture, larch plantation and farmland. Shotgun metagenomic and metatranscriptomic analyses of soil samples exposed to elevated or low H-2 concentration led to the identification of 45 genes encoding the large subunit of [NiFe] hydrogenases belonging to 8 distinct phyla. Our results indicate that despite significant sequencing effort, retrieved hydrogenase sequences are not in themselves adequate surrogates of H-2 oxidation activity in these soils. In fact, land-use exerted a greater influence than H-2 exposure on both hydrogenase gene distribution and expression though expression of certain genes responded to H-2. We argue that approaches relying on PCR/RT-PCR amplicon sequencing or quantification combined with physicochemical parameters are currently the best option to infer the activity of H-2-oxidizing bacteria and probably other specialist functional guilds with similar population size in soil.