Fig. 3.

Unrooted phylogenetic tree of β-Kdo GTs. Kdo-transferase domains in multidomain proteins were separated based on the conserved domains identified using the NCBI Conserved Domain Search, disordered regions in Phyre2 secondary structure models, and biochemical data (see SI Materials and Methods for details). Multiple sequence alignment was performed with MAFFT followed by curation in GBlocks (Fig. S2). The maximum-likelihood tree was calculated using PhyML. Bootstrap values above 50% are shown adjacent to nodes and are based on 1,000 replications. The accession numbers of the proteins are given in Table S1.