Author Response: Structural and Functional Properties of a Magnesium Transporter of the SLC11/NRAMP Family

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Members of the ubiquitous SLC11/NRAMP family catalyze the uptake of divalent transition metal ions into cells. They have evolved to efficiently select these trace elements from a large pool of Ca2+ and Mg2+, which are both orders of magnitude more abundant, and to concentrate them in the cytoplasm aided by the cotransport of H+ serving as energy source. In the present study, we have characterized a member of a distant clade of the family found in prokaryotes, termed NRMTs, that were proposed to function as transporters of Mg2+. The protein transports Mg2+ and Mn2+ but not Ca2+ by a mechanism that is not coupled to H+. Structures determined by cryo-EM and X-ray crystallography revealed a generally similar protein architecture compared to classical NRAMPs, with a restructured ion binding site whose increased volume provides suitable interactions with ions that likely have retained much of their hydration shell. Editor's evaluation This work elegantly fuses cryo-EM, X-ray crystallography, and in vitro transport experiments to describe the structural basis for functional diversity in the SLC11/NRAMP family of membrane transporters. This work identifies factors responsible for selectivity of classical NRAMPS for transition metal ions (Fe, Mn) and the NRMT clade for alkali metal ion (Mg). Although selectivity is much discussed in transport of divalent metal ions, this is an outstanding example of a study that gets to the bottom of the structural determinants governing this behavior. https://doi.org/10.7554/eLife.74589.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Divalent cations constitute important factors in numerous biological processes and they are thus essential nutrients that need to be imported into cells in the required amounts. From a chemical perspective, we distinguish the alkaline earth metals Ca2+ and Mg2+ from transition metal ions such as Fe2+ and Mn2+, which show distinct properties owing to their size and electronic structure. As major constituent of bone, Ca2+ has to be taken up by vertebrates in high amount. Although its largest part is immobilized within the body, free Ca2+ is present in the extracellular medium in low millimolar concentrations. In contrast, cytoplasmic Ca2+ serves as second messenger and its distribution in the cell thus requires tight control (Carafoli and Krebs, 2016). Mg2+ on the other hand is a ubiquitous ligand of nucleotides and nucleic acids (de Baaij et al., 2015; Maguire and Cowan, 2002). Despite the similar extracellular concentration, its abundance in the cytoplasm is several orders of magnitude higher compared to Ca2+. In their interaction with other molecules, these alkaline earth metal ions are coordinated by electronegative hard ligands, preferentially oxygens, which compensate their positive charges by either full or partial countercharges (Carafoli and Krebs, 2016; Payandeh et al., 2013). In contrast to Ca2+ and Mg2+, which both constitute minor elements in the human body, transition metals are trace elements that are found at much lower concentration, as they are required for specialized processes. Transition metals are distinguished by their incompletely filled d-orbitals. This property permits coordinative interactions with soft ligands containing free electron pairs and allows them to change their oxidation state, which makes them important cofactors in redox reactions. The transition metal ion Fe2+ plays a central role in oxygen transport and, together with Mn2+, it is a cofactor of enzymes catalyzing redox reactions. Due to their low abundance in the extracellular environment, the uptake of transition metals has thus to proceed with high selectivity to prevent competition with several orders of magnitude more abundant alkaline earth metals, which would prohibit their efficient accumulation. This challenge is overcome by specific transmembrane transport proteins, which catalyze the accumulation of transition metal ions inside cells. Among these proteins, members of the SLC11/NRAMP family play an important role. They are expressed in all kingdoms of life where they facilitate the transmembrane transport of different transition metal ions by a secondary active process that involves the cotransport of H+ as energy source (Courville et al., 2006; Nevo and Nelson, 2006). In animals, these proteins are used for the transport of Fe2+ (Montalbetti et al., 2013), whereas in prokaryotes the primary substrate is Mn2+ (Makui et al., 2000). Besides the transport of both transition metals, plant NRAMPs are also involved in detoxification processes by transporting noxious metal ions such as Cd2+ (Huang et al., 2020a). The transport properties of SLC11/NRAMP proteins have been investigated in numerous functional studies (Gunshin et al., 1997; Mackenzie et al., 2006; Makui et al., 2000), supported by the structures of prokaryotic transporters determined in different conformations (Bozzi et al., 2016b; Bozzi et al., 2019b; Ehrnstorfer et al., 2014; Ehrnstorfer et al., 2017). With respect to its fold, the family is distantly related to other secondary active transporters that include the amino-acid transporter LeuT (Ehrnstorfer et al., 2014; Yamashita et al., 2005), although there is little relationship on a sequence level, SLC11/NRAMP structures provide detailed insight into substrate preference and transport mechanisms. Characterized family members show an exquisite selectivity for transition metals over alkaline earth metal ions, but they poorly discriminate between transition metals (Ehrnstorfer et al., 2014; Gunshin et al., 1997). This is accomplished by a conserved ion binding site that is located in the center of the protein (Ehrnstorfer et al., 2014). Several residues coordinate the largely dehydrated metal ion with hard, oxygen-containing ligands, which would equally well interact with other divalent cations. Additionally, the site allows interaction with the free electron pair of the thioether group of a conserved methionine residue, which serves as soft ligand that permits coordinative interactions with transition metal- but not alkaline earth metal ions (Ehrnstorfer et al., 2014). The mutation of this methionine to alanine in one of the bacterial homologues had a comparably small effect on transition metal ion transport but instead converted Ca2+ into a transported substrate, which emphasizes the role of the residue in conferring ion selectivity (Bozzi et al., 2016a). Although, throughout kingdoms of life, the majority of SLC11/NRMP transporters share similar functional properties, there are few family members that were described to transport with distinct substrate preference. Among these family members are prokaryotic proteins, termed NRAMP-related magnesium transporters (NRMTs), that were identified as uptake systems for Mg2+. In the bacterium Clostridium acetobutylicum, the prototypic NRMT (CabNRMT) has permitted growth at limiting Mg2+ concentrations, even in absence of any alternative transport pathways (Shin et al., 2014). The altered substrate preference is remarkable since, owing to its small ionic radius and the consequent high charge density, the interaction of Mg2+ with its surrounding solvent is unusually strong, which makes the dehydration of the ion energetically costly and therefore requires unique features for its transport (Chaudhari et al., 2020; Maguire and Cowan, 2002). The uptake of Mg2+ into cells is thus accomplished by few selective transport systems (Payandeh et al., 2013) that include the bacterial proteins CorA (Lunin et al., 2006), CorC (Huang et al., 2021), MgtE (Hattori et al., 2007) and their eukaryotic homologues (Schäffers et al., 2018; Schweigel-Röntgen and Kolisek, 2014), the TRP channel TRPM7 (Huang et al., 2020b) and Mg2+-selective P-type pumps (Maguire, 1992). In the present study, we became interested in the mechanism of Mg2+ transport by SLC11 proteins and we thus set out to investigate the functional and structural basis of this process by combining transport experiments with structural studies. We identified a biochemically well-behaved NRMT, which allowed us to characterize its transport behavior using in vitro transport experiments. Structure determination by cryo-electron microscopy and X-ray crystallography revealed an expanded substrate binding site that binds the divalent cation as complex with most of its hydration shell, which permits selection against the larger Ca2+, which is not a transported substrate. Results Phylogenetic analysis of NRAMP transporters In light of the proposed altered substrate preference of certain prokaryotic SLC11 proteins, we became interested in the phylogeny of the family. For that purpose, we have identified numerous SLC11 homologues in BLAST searches of pro- and eukaryotic sequence databases using human DMT1, the prokaryotic Mn2+ transporters EcoDMT and ScaDMT, the putative NRAMP-related aluminium transporter (NRAT) from Orzya sativa and the proposed Mg2+ transporter CabNRMT as queries. In this way, sequences of 1100 eukaryotic and 447 prokaryotic NRAMP proteins and 745 prokaryotic NRMTs were aligned and subjected to phylogenetic characterization. The resulting phylogenetic tree displayed in Figure 1 shows a distributed organization with different family members clustering into separated clades. Branches of this phylogenetic tree group proteins from the animal kingdom including human DMT1 and NRAMP1, which both function as H+ coupled Fe2+ transporters (Forbes and Gros, 2003; Gunshin et al., 1997), prokaryotic homologues including proteins of known structure such as ScaDMT (Ehrnstorfer et al., 2014), EcoDMT (Ehrnstorfer et al., 2017) and DraNRAMP (Bozzi et al., 2019b), whose primary function is H+ coupled Mn2+ transport and plant proteins ascribed to be involved in the transport of various transition metal ions such as the toxic metal Cd2+ (Huang et al., 2020a). Within the plant kingdom, a small but separated branch encompasses putative Al3+ transporters classified as NRATs (Chauhan et al., 2021; Xia et al., 2010). Finally, a large and distant clade of the family containing homologues of the prototypic CabNRMT defines a group of transporters which, although clearly being part of the family, appear to have evolved to serve a different purpose (Figure 1B; Shin et al., 2014). The proposed functional distinction between different clades is manifested in the sequence of the substrate binding site that was identified in structural studies of prokaryotic SLC11 family members to coordinate divalent transition metals (Ehrnstorfer et al., 2014; Figure 1A, Figure 1—figure supplement 1). This site is defined by a signature containing a DxxN motif in the unwound center of α-helix 1 and an A/CxxM motif in the unwound center of α-helix 6. The signature is fully conserved in animal Fe2+ transporters, bacterial Mn2+ transporters and most transporters of plant origin, which all exhibit a strong selectivity against alkaline earth metal ions but poor discrimination between transition metal ions. Conversely, the motif is replaced by a DxxN-AxxT motif in NRATs and a DxxG-TxxA/T motif in NRMTs thus indicating that latter clades would exhibit a different substrate preference (Figure 1—figure supplement 1). Figure 1 with 1 supplement see all Download asset Open asset Phylogenetic analysis of the SLC11/NRAMP family. (A) Model of the consensus ion binding site of NRAMP transporters as obtained from the X-ray structure of ScaDMT (PDBID 5M95) in complex with Mn2+. Regions of α1 and α6 are shown as Cα-trace, selected residues coordinating the ion as sticks. (B) Phylogenetic tree of SLC11 residues with different clades of the family highlighted. Selected family members are indicated. The consensus sequence of the respective ion binding site of each clade is shown. Functional characterization of EcoDMT mutants To gain further insight into the presumed substrate preference, we investigated how mutations of ion binding site residues would alter the selectivity of a classical NRAMP protein by characterizing the functional properties of the bacterial Mn2+ transporter EcoDMT. To do so, we employed previously established proteoliposome-based transport assays with appropriate fluorophores to detect the uptake of specific ions into vesicles (Ehrnstorfer et al., 2017, Figure 2—figure supplement 1). All transport experiments were carried out in in presence of a 100-fold outwardly directed K+ gradient, which after addition of the ionophore valinomycin establishes a membrane potential of –118 mV, to enhance the sensitivity of the applied assays. Classical NRAMP proteins efficiently transport divalent transition metal ions with low micromolar affinity, while discriminating against alkaline earth metal ions such as Ca2+ and Mg2+ (Ehrnstorfer et al., 2017; Mackenzie et al., 2006). These functional properties are reflected in transport studies of EcoDMT, which employ the fluorophore calcein as efficient reporter to monitor micromolar concentrations of Mn2+ whereas it is not responsive to either Ca2+ or Mg2+ even at much higher concentrations. In our experiments, we find a concentration-dependent quenching of the fluorescence due to Mn2+ influx into vesicles containing the trapped fluorophore (Figure 2A and B, Figure 2—figure supplement 1A), which is neither inhibited by the addition of Ca2+ nor Mg2+ to the outside solution (Figure 2C and D). Latter would be expected if these ions would compete for the binding of Mn2+. As suggested from previous studies (Bozzi et al., 2016a; Ehrnstorfer et al., 2014), we expected a key role of the methionine (Met 235) in the binding site of the protein as major determinant of this pronounced selectivity. We thus initially investigated whether it was possible to alter the substrate preference of EcoDMT by shortening the methionine side chain in the mutant EcoDMT M235A. This construct was still capable of transporting Mn2+ with a threefold increase of its Km (Figure 2A, B, E and G), whereas now also Ca2+ became a permeable substrate as illustrated by the competition of the ion with Mn2+ transport (Figure 2F and G) and the direct detection of Ca2+ transport using the fluorophore Fura2 (Figure 2H, Figure 2—figure supplement 1B). A similar result was previously obtained for the equivalent mutation of the homologue DraNRAMP (Bozzi et al., 2016a). In contrast to Ca2+, we were not able to detect any interaction with Mg2+ (Figure 2I), indicating that the energetic requirements to transport this smaller divalent cation are higher than for Ca2+. Figure 2 with 1 supplement see all Download asset Open asset Functional characterization of EcoDMT mutants. (A) EcoDMT-mediated Mn2+ transport into proteoliposomes. Data display mean of three experiments from three independent reconstitutions. (B) Mn2+ concentration dependence of transport. Initial velocities were derived from individual traces of experiments displayed in (A), the solid line shows the fit to a Michaelis–Menten equation with an apparent Km of 13.5 μM. (C) Mn2+ transport in presence of Ca2+. Data display mean of seven experiments from two independent reconstitutions, except for the measurement without divalent ions (0/0, green, mean of two experiments). (D) Mn2+ transport into EcoDMT proteoliposomes in presence of Mg2+. Data display mean of 3 (0/0, green), 4 (50/0, blue) and nine experiments (50/500, orange and 50/2000, red) from three independent reconstitutions. (E) Mn2+ transport into proteoliposomes containing the EcoDMT mutant M235A. Data display mean of eight experiments from three independent reconstitutions. (F) Inhibition of Mn2+ transport in the mutant M235A by Ca2+ (3 experiments from three independent reconstitutions). (G) Mn2+ concentration dependence of transport into M235A proteoliposomes. For Mn2+, initial velocities were derived from individual traces of experiments displayed in (E), for Mn2+ in presence of 2 mM Ca2+, data show mean of three experiments from three independent reconstitutions. The solid lines are fits to a Michaelis–Menten equation with apparent Km of 41 μM (Mn2+) and 107 uM (Mn2+ in presence of 2 mM Ca2+). (H) Ca2+-transport into M235A proteoliposomes assayed with the fluorophore Fura-2. Data display mean of tree experiments from three independent reconstitutions. (I) Mn2+ transport in presence of Mg2+. Data display mean of 6 experiments from three independent reconstitutions. (J) Mn2+ transport into proteoliposomes containing the EcoDMT triple mutant M235A/N54G/A232T. Data display mean of three experiments from three independent reconstitutions. (K) Inhibition of Mn2+ transport in the mutant M235A/N54G/A232T by Mg2+ (six experiments from three independent reconstitutions). (L) Mn2+ concentration dependence of transport into M235A/N54G/A232T proteoliposomes. For Mn2+, initial velocities were derived from individual traces of experiments displayed in (J), for Mn2+ in presence of 2 mM Mg2+, data show mean of three (without Mg2+, red) and six experiments (with Mg2+, blue) from three independent reconstitutions. The solid lines are fits to a Michaelis–Menten equation with apparent Km of 31 μM (Mn2+) and 38 µM (Mn2+ in presence of 2 mM Mg2+). (A, C, D, E, F, I, J, K) Uptake of Mn2+ was assayed by the quenching of the fluorophore calcein trapped inside the vesicles. (A, C, D, E, F, H, I, J, K) Averaged traces are presented in unique colors. Fluorescence is normalized to the value after addition of substrate (t = 0). Applied ion concentrations are indicated. (A–L), Data show mean of the indicated number of experiments, errors are s.e.m. Finally, we attempted to confer Mg2+ transport properties to EcoDMT by creating the triple mutant M235A/N54G/A232T, which converts its entire binding site to the residues found in prokaryotic NRMTs. As for the mutant M235A, this construct folded into a stable protein that retained its capability of Mn2+ transport while remaining unable to interact with Mg2+ (Figure 2J–L). In summary, the mutation of the binding site residues of EcoDMT towards an NRMT was well tolerated with the resulting constructs retaining their capability for Mn2+ transport. This suggests that the structural and energetic requirements for interactions with this transition metal are lower than for alkaline earth metals. We also confirmed the role of the binding site methionine as an important determinant for the counterselection of Ca2+, which due to its much higher concentration would outcompete Mn2+ transport in a physiological environment. However, neither the replacement of the methionine nor the conversion of the binding site to residues observed in NRMTs was sufficient to facilitate Mg2+ transport, demonstrating that structural features beyond the immediate binding site residues are required to transport this small divalent cation. Consequently, we decided to turn our attention towards the characterization of prokaryotic NRMTs to gain further insight into Mg2+ transport by SLC11 homologues. Functional characterization of EleNRMT Initially, our attempts toward the characterization of NRMTs were focused on the identification of a suitable candidate for structural and functional studies. Towards this end, we cloned and investigated the expression properties of 82 representative homologues selected from the pool of 745 sequences of putative NRMTs identified in BLAST searches (Figure 1B). Within this subset, we were able to single out a candidate from the bacterium Eggerthella lenta termed EleNRMT with promising biochemical properties (Figure 3—figure supplement 1). After expression and purification in the detergent n-dodecyl-beta-D-Maltoside (DDM), we found this protein to elute as monodisperse peak from a size exclusion chromatography column. (Figure 3—figure supplement 2A). Although much better behaved than the other investigated homologues, EleNRMT was still comparably instable in detergent solution and difficult to reconstitute into liposomes. We thus attempted to improve its stability by consensus mutagenesis (Cirri et al., 2018) by converting 11 positions identified in sequence alignments to their most abundant residues in NRMTs to create the construct EleNRMTts (Figure 3—figure supplement 1A). This approach improved the thermal stability of the monomeric protein and the efficiency of reconstitution without affecting its functional properties (Figure 3, Figure 3—figure supplement 2B-D). Due to the sensitivity of the assay and the fact that binding site mutants of EcoDMT retained their ability to transport Mn2+, we initially investigated transport properties of EleNRMT for this divalent transition metal ion and found robust concentration-dependent activity that saturates with a Km of around 120 µM for either WT or the thermostabilized mutant (Figure 3A, B, D and E). For both constructs, we found Mg2+ to inhibit Mn2+ transport in a concentration-dependent manner, thus indicating that Mg2+ would compete with Mn2+ for its binding site (Figure 3C and F). We subsequently quantified binding of either ion to detergent solubilized protein by isothermal titration calorimetry and found a Kd of around 100 µM for Mn2+, which corresponds well with the measured Km of transport of the same ion and a somewhat lower affinity of around 400 µM for Mg2+ (Figure 3G and H, Figure 3—figure supplement 2E, F). Together, transport and binding experiments demonstrate the capability of EleNRMT to interact with Mg2+ thus providing evidence that it might indeed catalyze Mg2+ uptake into bacteria. Figure 3 with 2 supplements see all Download asset Open asset Transport properties of EleNRMT. (A) EleNRMT mediated Mn2+ transport into proteoliposomes. Data display mean of three experiments from two independent reconstitutions. (B) Mn2+ concentration dependence of transport. Initial velocities were derived from individual traces of experiments displayed in (A), the solid line shows the fit to a Michaelis–Menten equation with an apparent Km of 119 μM. (C) Mn2+ transport in presence of Mg2+. Data display mean of 3 (0/0, green), 6 (50/0, blue), 5 (50/500, orange), and 9 (50/2000, red) experiments from two independent reconstitutions. (D) Mn2+ transport into proteoliposomes mediated by the thermostabilized mutant EleNRMTts. Data display mean of four experiments from two independent reconstitutions, except for the measurement with 25 μM Mn2+ (mean of three experiments). (E) Mn2+ concentration dependence of transport. Initial velocities were derived from individual traces of experiments displayed in (D), the solid line shows the fit to a Michaelis–Menten equation with an apparent Km of 124 μM. (F) Mn2+ transport in presence of Mg2+. Data display mean of 4 (0/0, green), 9 (50/0, blue), 11 (50/500, orange), and 8 (50/2000, red) experiments from two independent reconstitutions. A, C, D, F. Uptake of Mn2+ was assayed by the quenching of the fluorophore calcein trapped inside the vesicles. Averaged traces are presented in unique colors. Fluorescence is normalized to the value after addition of substrate (t = 0). Applied ion concentrations are indicated. (A–F), Data show mean of the indicated number of experiments, errors are s.e.m. (G–H), Binding isotherms obtained from isothermal titrations of Mn2+ (G) and Mg2+ (H) to EleNRMTts. The data shown for two biological replicates per condition was fitted to a model assuming a single binding site with the binding isotherm depicted as solid line. Errors represent fitting errors. To further characterize Mg2+ transport by EleNRMT, we have quantified Mn2+ uptake into proteoliposomes in presence of Mg2+ and found a strong interference (Figure 4A and B). By directly assaying Mg2+ with the selective fluorophore Magnesium Green, we observed a concentration-dependent increase of the transport rate (Figure 4C, Figure 2—figure supplement 1C), with the lack of saturation at high substrate concentrations, reflecting the poor affinity of the dye for the ion. Since our experiments have confirmed Mg2+ as substrate of EleNRMT, we wondered whether the protein would also transport Ca2+ and thus investigated the inhibition of Mn2+ uptake by Ca2+ and the import of Ca2+ into proteoliposomes monitored with the fluorophore Fura-2 (Figure 4D and E). However, in contrast to Mg2+, we did in no case find strong evidence for the recognition of this larger alkaline earth metal ion, which has become a transported substrate for EcoDMT mutants lacking a conserved methionine in the binding site (Figure 2F–H). Latter finding is remarkable since the presumed metal ion binding site only consists of hard ligands that would also coordinate Ca2+ and it thus emphasizes the presence of distinct structural features of the pocket facilitating Mg2+ interactions (Figure 3—figure supplement 1B). Figure 4 with 1 supplement see all Download asset Open asset Ion selectivity and substrate coupling. (A–E) Metal ion transport by EleNRMTts. (A) EleNRMTts mediated Mn2+ transport into proteoliposomes in presence of 2 mM Mg2+. Data display mean of eight experiments from three independent reconstitutions, except for the measurement with 200 μM Mn2+ (mean of seven experiments). (B) Mn2+ concentration dependence of transport into EleNRMTts proteoliposomes. For Mn2+, data are as displayed in Figure 3E, for Mn2+ in presence of 2 mM Mg2+, initial velocities were derived from individual traces of experiments displayed in (A). The solid lines are fits to a Michaelis–Menten equation. (C) Mg2+-transport into EleNRMTts proteoliposomes assayed with the fluorophore Magnesium Green. Data display mean of 5 (0 μM Mg2+), 8 (800 μM and 1250 μM Mg2+), 4 (2500 μM Mg2+), and 3 (5000 μM Mg2+) experiments from two independent reconstitutions. (D) Mn2+ transport in presence of Ca2+. Data display mean of 4–11 experiments from three independent reconstitutions. (A, D) Uptake of Mn2+ was assayed by the quenching of the fluorophore calcein trapped inside the vesicles. (E) Ca2+-transport into EleNRMTts proteoliposomes assayed with the fluorophore Fura-2. Data display mean of 3 (0 μM Ca2+), 12 (200 μM Ca2+) and 9 (2000 μM Ca2+) experiments from 2 independent reconstitutions. The small increase of the fluorescence at high Ca2+ concentrations likely results from non-specific leak into proteoliposomes. (F–H) Assay of H+ transport with the fluorophore ACMA. Experiments probing metal ion coupled H+ transport into proteoliposomes containing EleNRMTts upon addition of Mn2+ (eight experiments from two independent reconstitutions) (F) and Mg2+ (eight experiments from two independent reconstitutions) (H) do not show any indication of H + transport. Mn2+ coupled H+ transport into EcoDMT proteoliposomes (three experiments from three independent reconstitutions and for the negative control [no protein, 400 μM Mn2+], 4 measurements) is shown for comparison (I). A, C-H Averaged traces are presented in unique colors. Fluorescence is normalized to the value after addition of substrate (t = 0). Applied ion concentrations are indicated. (A–H), Data show mean of the indicated number of experiments, errors are s.e.m. After confirming the capability of EleNRMT to specifically transport Mn2+ and Mg2+, we were interested whether the uptake of metal ions would be coupled to H+ as observed for previously characterized transition metal ion transporters of the family. To this end we employed the fluorophore ACMA to monitor pH changes that are induced by metal ion symport (Figure 4F–H). Whereas such concentration-dependent changes are readily observed for EcoDMT (Figure 4H), there was no acidification of proteoliposomes containing EleNRMT as a consequence of either Mg2+ or Mn2+ uptake (Figure 4F and G), thus suggesting that metal ion transport by EleNRMT is not coupled to H+. Collectively our functional experiments confirm the role of EleNRMT to function as uncoupled Mg2+ transporter. Due to the high-sequence conservation in this subclade of the SLC11 family, it is safe to assume that this property would likely also be shared by other NRMTs. Structural characterization of EleNRMT Following the characterization of the transport properties of EleNRMT, we became interested in the molecular determinants underlying its distinct selectivity and thus engaged in structural studies by X-ray crystallography and cryo-EM. Since the monomeric protein, with a molecular weight of only 47 kDa and lacking pronounced domains extending from the membrane, is itself not a suitable target for cryo-EM, we initially attempted structural studies by X-ray crystallography but were unable to obtain crystals, despite extensive screening. To enlarge the size of the protein and facilitate crystallization, we generated specific nanobodies by immunization of alpacas with purified EleNRMTts followed by the assembly of a nanobody library from blood samples and selection by phage display. These experiments allowed us to identify three distinct binders recognizing the protein termed Nb1-3EleNRMT (short Nb1-3) (Figure 4—figure supplement 1A). Despite the increase of the hydrophilic surface, which was frequently found to facilitate crystallization, all attempts to obtain crystals of EleNRMTts in complex with a single nanobody were unsuccessful. Similarly, efforts toward the structure determination of the same complexes by cryo-EM did not permit a reconstruction at high resolution. We thus attempted to further increase the size of the complex and found that Nb1 and either one of the two closely related nanobodies Nb2 or Nb3 were able to bind concomitantly to the protein to assemble into a ternary complex (Figure 4—figure supplement 1B-D). This complex sho