What They Saw
They knocked out different genes (or the whole operon) in the hyp operons of A. vinelandii CA by inserting resistance + lacZ cassettes, then tested these strains for hydrogen oxidation and expression of the genes (via beta-galactosidase activity). The medium they used had a lot more trace elements than typical Burk medium, including nickel.
They found that when lacZ was inserted in the same direction as the gene, they saw expression in all cases, whether fixing nitrogen or not, but if it was inserted in the opposite direction, they didn't. So apparently the genes are expressed to some extent even when not fixing nitrogen.
When they measured hydrogen oxidation, there was about 6x more when fixing nitrogen though, in the wild-type, and none in the mutants. The growth rates of the mutants were similar to the wild-type though, or so they claim without reporting any details.
Finally they tried growing the strains with extra added nickel, because the hydrogenase is a nickel-containing enzyme. This didn't have much effect on the wild-type, but the hypB mutant actually showed some hydrogenase activity in nitrogen-fixing conditions with the extra nickel, and activity was higher with more nickel added.
What This Means
It seems like extra added nickel can substitute for the lack of HypB, so maybe the enzyme has a role in nickel cofactor processing somehow. This makes sense considering the multiple histidines it contains. It's interesting that nickel didn't help activity in non-fixing conditions; maybe there are multiple hydrogenases, active in different conditions, and hypB is required for all but nickel doesn't help some of them.