What They Saw
The sequences were similar to A. vinelandii's hoxKG structural genes. They tried knocking each out, then measuring hydrogen oxidation (with methylene blue) and hydrogen production (with methyl viologen). As expected, hydrogen oxidation in mutants was no higher than negative controls. Surprisingly, they did see hydrogen production in the some of the different mutants with a strong electron donor, but it was less than in the wild-type. Only the mutant with an insertion very close to the start of the hupS gene had no hydrogen production.
What This Means
It seems like a fragment of HupS is sufficient to produce hydrogen with a strong electron donor, but not as much as with both HupS and HupL completely intact.