In the D. vulgaris structure, the residues making contact between heterodimers at the central interface are two proline residues (P368B1/B2) that stack on each other (Table 1, Figure 6A), whereas the interacting residues in the MV2-Eury homology model based on the D. vulgaris structure are two arginine residues (R337B1/B2) detected in the EVCoupling analysis as possible monomeric FPECs (Table 1, Figure 6B). This central interfacial residue is found to be an asparagine in 33.5% of the sequences and a serine in 7% of the sequences. In 42% of the sequences in the Dsr multiple sequence alignment (MSA) used by the EV Couplings server, the C-terminus is truncated prior to this residue at the central interface, which may correspond to sequencing errors or inclusion of partial sequences in this database. Note that, in the D. vulgaris structure, the B1 and B2 helices at the central interface swap positions with respect to the A. fulgidus structure (Figure 1). The aromatic residues, F365B1 and Y412A2, in the D. vulgaris heme road are switched with respect to Y348B1 and F394A2 in A. fulgidus . Like the MV2-Eury heme road, the third aromatic residue in A. fulgidus, F317A1 is replaced by a residue from the B1 subunit, I190B1.
Examination of the heme road residues in the structural alignment of the MV2-Eury homology and AF2 models reveals that they are slightly shifted between the two models, with S167 of the B1/2 subunits in the AF2 model exhibiting a steric clash between models with two of the Fe-S centers in the homology model. The two residues constituting the central interface between heterodimers in MV2-Eury B1/2 subunits, R337B1-R337B2, are superimposed at the level of the backbone between the homology and AF2 models, but the side chains point away from each other in the AF2 model, while they interact closely with each other in the homology model (Figure S3). The positioning of the A-subunit C-terminal heme road residue, N369A1/2, in MV2-Eury is quite similar between the homology and AF2 models, despite the very different positioning of the C-terminal helix (Figure 4). The three residues in each heterodimer comprising the heme road on either side of the central interface are not in direct contact, supporting the notion that the co-variance is not uniquely structural in nature. Results of the evolutionary coupling using different subunits from different homologues as bait supports the notion that at least one pair of these residues, T351B1-N393A2 and vice versa in A. fulgidus , is truly co-varying. The positioning of N180B1/2 in interaction with the siroheme and relatively close to N393A2/1 is suggestive that this constitutes a second co-varying pair. Based on these observations, we hypothesize that the heme road represents a pathway of cooperative communication between active sites within the DsrAB heterotetramer. The aromatic residues packed between them (Figure 5D) could help modulate information transfer in this putative allosteric route. Alternatively, given the relatively short distances between the heme road residues and the intervening aromatic residues, we cannot rule out that this represents a pathway for electron transfer between heterodimer active sites.