What They Saw
Running proteins on a gel testing for SOD activity, they saw two bands: one was Fe-SOD and the other CuZnSOD (which sits in the periplasm). They tried knocking out sodB from A. vinelandii by introducing a kanamycin resistance cassette, and isolated a kan-resistant strain, but it appeared to have two copies of sodB (only one of which was knocked out). They tried increasing the concentration of kanamycin (presumably to force the strain to have multiple copies of the resistance gene), and got one that grew slowly at 100x more kanamycin than I use, but they couldn't get rid of the SOD. Seems like it's essential.
This also supports the idea that A. vinelandii can have multiple copies of its chromosome, since they saw multiple PCR products from the same locus, with and without the resistance marker. The genome only has one copy of sodB, so there must be multiple genome copies.