Figure 4.

Neighbor-joining tree of ferredoxin nitrite reductase (NirA) sequences obtained from KEGG [24], RefSeq [25], and 2 novel clones (LostP-NirA-8 and -25). All sequences were truncated to the segment between the binding sites of primer pair arc-NirA, aligned by a ClustalW algorithm with BLOSUM cost matrix, and tree construction followed a Jukes-Cantor model with bootstrap resampling (1000 replicates). The eukaryotic NirA sequence of Aspergillus flavus was used for an outgroup. Branch labels indicate bootstrap support in %; branches with <50% support were collapsed.