What They Saw
They already knew about the structural genes, hoxKG. So they got some overlapping fragments containing those and sequences downstream of them, and sequenced them. They found 5 new open reading frames (ORFs), ORF3-7. They named these hoxZMLOQ. There was also part of an eighth at the end. All of them were homologous to hydrogenase genes in other organisms.
Then they knocked out each gene by inserting kanamycin or kan+lacZ cassettes, though probably some of these insertions affected other genes (polar effects) so it wasn't possible to study each gene individually. Each insertion abolished hydrogen oxidation. All 5 new ORFs seem to be involved in subunit processing, as the unprocessed form was present in higher proportion (or alone) in their knockouts.
The knockouts didn't seem to be impaired in growth, even when fixing nitrogen, though they might show more effect in certain conditions (like carbon limitation).
HoxZ seems to be a membrane protein, possibly a cytochrome, so part of the electron transport chain. HoxL seems like it might be involved in metal-binding somehow.
What This Means
Knowing more of the sequence of the operon helps to study it. The whole thing, hox and hyp, is pretty big, kinda surprising for an enzyme that doesn't seem to help the organism in some conditions. Probably in nature, the likelihood of encountering carbon-limited circumstances is much higher, so having an uptake hydrogenase is important.