The metal binding site composition of the cation diffusion facilitator protein MamM cytoplasmic domain impacts its metal responsivity

The cation diffusion facilitator (CDF) is a conserved family of divalent d-block metal cation transporters that extrude these cations selectively from the cytoplasm. CDF proteins are composed of two domains: the transmembrane domain, through which the cations are transported, and a regulatory cytoplasmic C-terminal domain (CTD). It was recently shown that the CTD of the CDF protein MamM from magnetotactic bacteria has a role in metal selectivity, as binding of different metal cations exhibits distinctive affinities and conformations. It is yet unclear whether the composition of the CTD binding sites can impact metal selectivity and if we can manipulate the CTD to response to other non-native metals in CDF proteins. Here we performed a mutational study of the model protein MamM CTD, where we exchanged the native metal binding residues with different metal-binding amino acids. Using X-ray crystallography and Trp-fluorescence spectrometry, we studied the impact of these mutations on the CTD conformation in the presence of non-native metals. Our results reveal that the incorporation of such mutations alters the domain response to metals in vitro, as mutant forms of the CTD bind metals differently in terms of the composition of the binding sites and the CTD conformation. Therefore, the results demonstrate the direct influence of the CTD binding site composition on CDF proteins structure and hence, function, and constitute a first step for rational design of MamM for transporting different metals in vivo.


Supplementary Results and Discussion
In the main text we focused on the analyses of the differences between the ability of Zn 2+ , Ni 2+ and Mn 2+ to be bound by different mutants (see main text Results and Discussion). Here we give detailed analyses of Cu 2+ and Cd 2+ binding to the varied mutants.
MamM CTD mutants binding to Cu 2+ : For this study we solved the crystal structures of MamM CTD D249H and D249N with Cu 2+ , while we previously solved the structure of the Cu 2+ -bound WT CTD 1 .
The WT bound structure shows tighter conformation compared to the WT apo form ( Figure S5A), with Cu 2+ ions bound in the central binding site (one ion by H285 from both monomers) and peripheral binding sites (in each site, one ion by H264 from one monomer and a water molecule that bridges the E289 residue from the second monomer) ( Figure S5B). Both D249H and D249N Cu 2+ -bound structures exhibit the same conformation as the WT apo form, with the Cu 2+ only bound by the H264 residues and water molecules ( Figure S5C). Although in the WT bound structure the D249 does not participate in the chelation of the Cu 2+ ions, it appears from the D249H and D249N Cu 2+ -bound structures that this residue is important for the correct metal binding and conformational change. The bound Cu 2+ by H264 in the D249N/H structures again hint at the importance of this residue for the chelation of the Cu 2+ ions.
Even if the H264 binding does not lead to conformational changes, it might act as an intermediate binding site. This is also supported by the Trp-fluorescence results of H264E, that show the least quenching of all mutants. In contrast to the crystal structures, the Trp-fluorescence results of the WT and the D249N/H mutants ( Figure 2E, lower panel) show that D249H has similar quenching to the WT, while the D249N exhibits much less quenching. Hence, the asparagine in position 249 seems to inhibit the binding to the H285 residues, thereby causing less quenching, while the histidine does not. This suggests that in the case of D249H, the crystal structure does not accurately represent the state in solution. Cu 2+ was shown to be bound almost exclusively by histidine in crystal structures 2 , and in the WT Cu 2+ -bound structure it also appeared that the chelation by the histidine residues mostly contributes to the conformation stability. Hence, we propose that in contrast to the N249 residue in the D249N mutants, the H249 residue contributes to the stabilization of the Cu 2+ binding in the central site and to the closed conformation, although not observed in the crystal structure. The Trp-fluorescence results of D249H-H285D show also the same quenching pattern as in the WT and D249H, suggesting that indeed a histidine residue in the central site is important for proper chelation and conformational change.
MamM CTD mutants binding to Cd 2+ : For this study we solved the crystal structures of MamM CTD D249H, D249E, D249N, H285D, H264E (two different SGs), H264E-E285H and E289D with Cd 2+ ; we previously solved the structure of the Cd 2+ -bound WT CTD 1 . Only two Cd 2+ -bound structures exhibit tighter conformation than the WT apo structure; one of the H264E structures and E289D ( Figure   S6A). Both structures exhibit the same conformation, but different binding site. While the H264E structure exhibits the same binding site as the H264E Zn 2+ -bound structure (H285 from one monomer and H236 and E289 from the second monomer), the E289D exhibits a unique Cd 2+ -pair binding site, as discussed in detail in the main text. All other MamM CTD constructs with Cd 2+ , including the WT, adopt the same conformation as the WT apo, and the Cd 2+ is bound only in the central binding site by the residues in positions 249, 285 or both ( Figure S6B). The exact Cd 2+ coordination depends on the substitution identity. In most mutants there is only one Cd 2+ ion bound to each monomer, while in others (WT and H264E) there are two ions that populate the binding site, but in half occupancy, suggesting that only one ion is effectively bound in them as well, by one of either binding sites. Furthermore, in some mutants the Cd 2+ is chelated by H213 and/or E215 from a non-biological monomer from an adjacent unit, which helps to form stable coordination of the Cd 2+ ions in the crystal form. Interestingly, in all structures where H285 residue exists it is involved in the chelation of the Cd 2+ ions, except in D249H where the H249 rather than H285 residue binds the ion. In the H285D mutant, the aspartate at position 249 binds the ion, overall showing that both the 249 and 285 positions and both histidine and aspartate residues can bind the Cd 2+ ions effectively. Previous PELDOR results show that the WT CTD binding to Cd 2+ lead to a tight rigid conformation, similar to that of the WT CTD with Zn 2+ and Cu 2+ , and in contrast to the Cd 2+ -bound crystal structure 1 . This indicates that for Cd 2+ , the crystal structures do not necessarily represent well the structures in solution. In some mutants, Cd 2+ binding might not lead to conformational changes (as discussed below). In mutants that do no exhibit conformational changes in the Cd 2+ -bound structures but do in solution, it is possible that Cd 2+ binding to each monomer prevents the tighter association of the monomers due to the large size of the Cd 2+ ions. Such an ion would not afford correct movement and stable tight conformation due to charge repulsion and steric interference. When considering the Trp-fluorescence results ( Figure 2B, top panel), one can see that in solution the conformational changes occur, with each construct exhibiting different conformation. For example, H264E shows no blue-shift but one of its crystal forms showed tighter conformation, suggesting that H264E Cd 2+ -bound structure, that exhibits no conformational change, is the conformation in solution (as discussed in the main text). E289D, whose crystal structure shows tighter conformation, shows a smaller blue shift compared to the WT. In this case, since: (1) The E289D-Cd 2+ and the H264E-Zn 2+ structures exhibit similar dimerization, (2) The addition of Cd 2+ to E289D and the addition of Zn 2+ to H264E result in smaller blue-shift as compared to their addition to the WT CTD, and (3) Zn 2+ -bound and Cd 2+ -bound conformations in solution are similar for the WT CTD (as evident by PELDOR 1 ), we propose that the crystal structure of E289D-Cd 2+ also represents well the structure in solution. D249H shows the largest blue-shift, suggesting that the binding to position 249 exclusively (this is the only form where the ion was bound only at this position) impacts more on the conformational changes or on the Trp signal (as position 249 is closer to W247 than position 285).  Values in parentheses are for the highest resolution shell.

Supplementary Tables
One crystal was used per dataset.
Data was collected at 100K.
b Resolution is given for best axis, as anisotropic data reduction in other axes resolution is worse.
c P-Preferred region, A-Allowed region, and O-outliers.     . Both structures exhibit similar conformation tighter than the WT apo, however the number of bound Cd 2+ ions per dimer and the binding sites differ between the two mutants (for E289D, two residues from a third monomer also participate in the chelation of each Cd 2+ -pair, see