|
Figure 1.
Functional expression of wild-type human and mouse ferroportin (Fpn) and mouse H32R-Fpn EGFP fusion proteins in Xenopus laevis oocytes. A: live-cell imaging of a control oocyte and oocytes expressing wild-type human ferroportin (hFpn), wild-type mouse ferroportin (mFpn), and mouse H32R-Fpn EGFP fusion proteins by using the confocal laser-scanning microscope. Each image is of a separate oocyte with an optical slice ≈11.4 µm roughly bisecting the oocyte. All images were collected using identical settings (pinhole, intensity, and gain). Scale bars = 0.2 mm. B: radiotracer metal (*M) efflux activity in control oocytes and oocytes expressing mFpn or H32R, comparing iron and manganese transport. Data are means, SD (n = 9, 12, 11, 10, 9, 12). Two-way ANOVA revealed an interaction (P < 0.001). Within 55Fe, mFpn differed from control (P < 0.001) but H32R did not (P = 0.97). Within 54Mn, neither mFpn (P = 0.09) nor H32R (P = 0.98) differed from control. C: 54Mn efflux as a function of the oocyte 54Mn injectate concentration (50 nL volume) in control oocytes or oocytes expressing mFpn (control, n = 12, 12, 8, 10 from low to high concentration; mFpn, n = 9, 11, 9, 8). Two-way ANOVA revealed a main effect of injectate 54Mn concentration (P < 0.001) but not of transporter expression (P = 0.78) and no interaction (P = 0.58).
|
|
Figure 2.
Metal substrate profile of mouse and human ferroportin. We assayed radiotracer metal (*M) efflux for each of six metals individually compared against Fe as the standard in control oocytes and oocytes expressing hFpn or mFpn. Oocytes were injected with 50 nL of 5 μM *M except as noted. Two-way ANOVA, for A and C–F, revealed an interaction (P < 0.001) and multiple pairwise comparisons vs. control confirmed in all cases that 55Fe efflux activity in hFpn and mFpn differed from control (P < 0.001). In B, two-way ANOVA revealed no interaction (P = 0.049). When *M efflux (other than 55Fe) differed from control for either hFpn or mFpn, we used a nested two-way analysis (excluding control) to test whether hFpn and mFpn differed in their *M efflux activity relative to 55Fe efflux activity. A: 109Cd efflux activity. Multiple pairwise comparisons vs. control (n = 11, 12): hFpn (n = 11, 12) and mFpn (n = 11, 11) did not differ from control (P ≥ 0.99). B: 57Co efflux activity. Multiple pairwise comparisons vs. control (n = 10, 9): hFpn (n = 10, 11) and mFpn (n = 12, 10) 57Co efflux activity differed from control (P < 0.001). Nested two-way analysis, no interaction (P ≥ 0.99). C: 64Cu efflux activity. Multiple pairwise comparisons vs. control (n = 10, 12): hFpn (n = 11, 11) and mFpn (n = 11, 8) did not differ from control (P ≥ 0.45). D: 54Mn efflux activity. Multiple pairwise comparisons vs. control (n = 9, 9): hFpn (n = 11, 12) and mFpn (n = 11, 9) did not differ from control (P ≥ 0.09). E: 63Ni efflux activity. Multiple pairwise comparisons vs. control (n = 10, 12): hFpn (n = 8, 11) and mFpn (n = 12, 8) 63Ni efflux activity differed from control (P < 0.001). Nested two-way analysis, no interaction (P = 0.40). F: 65Zn efflux activity (all groups: injectate [*M] = 50 µM; 3 days after RNA injection). Multiple pairwise comparisons vs. control (n = 12, 12): hFpn (n = 11, 12) and mFpn (n = 10, 11) 65Zn efflux activity differed from control (P < 0.001). Nested two-way analysis, no interaction (P = 0.16). hFpn, human ferroportin; mFpn, mouse ferroportin.
|
|
Figure 3.
Mouse ferroportin is activated by extracellular Ca2+. A: 55Fe efflux activity in the presence or absence of extracellular Ca2+ (0 mM Ca2+ solution contained 1 mM EGTA) in control oocytes (n = 12, 10) and oocytes expressing human Fpn (n = 12, 11) or mouse Fpn (n = 12, 10). Two-way ANOVA revealed an interaction (P < 0.001), whereas a nested two-way analysis comparing human and mouse Fpn revealed no interaction (P = 0.51). B: substitution of Ca2+ by other alkaline-earth metals. 55Fe efflux activity was measured in the presence of 3 mM magnesium (+Mg2+) or 1 mM Mg2+ plus 2 mM calcium (+Ca2+), strontium (+Sr2+), or barium (+Ba2+) in control oocytes (black, n = 23, 24, 24, 21) and oocytes expressing mouse Fpn (gray, n = 26, 26, 27, 25). Two-way ANOVA revealed an interaction (P < 0.001). Within individual conditions, mFpn differed from control in +Ca2+ and +Sr2+ (P < 0.001) but not in +Mg2+ (P = 0.10) or +Ba2+ (P = 0.032). Within mFpn, +Mg2+ and +Ba2+ differed from +Ca2+ (P < 0.001) but +Ca2+ and +Sr2+ did not differ (P = 0.54). hFpn, human ferroportin; mFpn, mouse ferroportin.
|
|
Figure 4.
Ferroportin-mediated iron-transport activity is pH-sensitive. 55Fe efflux activity as a function of extracellular pH (pHo) in control oocytes (gray, n = 19, 24, 11, 11, 17, 39, 20, 10, 35 low to high pH) and oocytes expressing human Fpn (green, n = 10, 11, 10, 12, 11, 12, 10) or mouse Fpn (blue, n = 16, 37, 34, 37, 35, 8, 52, 34, 16, 42). Data for human and mouse Fpn were fit by Eq. 2 and the results are described in Table 1. The solid lines represent the fit functions for human Fpn (green) and mouse Fpn (blue). Data for control oocytes could not be fit by Eq. 2 (r2 = 0.20, P = 0.75) nor by a linear function (r2 = 0.01, P = 0.83). Fpn, ferroportin.
|
|
Figure 5.
Mouse and human ferroportin are responsive to human hepcidin. 55Fe efflux activity in control oocytes (n = 12, 11) and in oocytes expressing human (n = 12, 10) or mouse Fpn (n = 12, 11) (5 days after RNA injection) untreated or pretreated 30 min with 10 µM human hepcidin. Two-way ANOVA revealed an interaction (P < 0.001); a nested two-way analysis within human and mouse Fpn revealed main effects of protein expression (P = 0.006) and of hepcidin treatment (P < 0.001) but no interaction (P = 0.31). Fpn, ferroportin.
|
|
Figure 6.
Effect of Na+ replacement on ferroportin-mediated iron transport activity. 55Fe efflux was assayed in the presence of 100 mM extracellular NaCl or 100 mM choline chloride in control oocytes (n = 12, 10) and oocytes expressing human Fpn (n = 12, 12) or mouse Fpn (n = 12, 13). Two-way ANOVA revealed a main effect of transporter expression (P < 0.001) but not of Na+ replacement (P = 0.85), and no interaction (P = 0.92). Fpn, ferroportin.
|
|
Figure 7.
Ferroportin is not an ATPase pump. We used the TargetATPSite and NSitePred algorithms in silico to search for predicted consensus ATP-binding sites in the human Fpn (A) and in the mouse Fpn (B) sequences. We used the recommended binary threshold probabilities of 0.7 for TargetATPSite (blue line) and 0.09 for NSitePred (green line). C: 55Fe efflux activity in control oocytes (n = 16, 14, 16, 17 per group) and oocytes expressing human Fpn (n = 17, 17, 17, 18) or mouse Fpn (n = 18, 18, 16, 15) that were untreated or treated with one of several metabolic inhibitors. Oocytes were both pretreated for 30 min with the metabolic inhibitors (10 µM rotenone, 4 µM antimycin A, and 12.6 µM oligomycin) and treated with the same inhibitors throughout the 30-min efflux assay. All solutions contained 0.1% DMSO. Two-way ANOVA revealed a main effect of transporter expression (P < 0.001) but not of inhibitor treatment (P = 0.021), and no interaction (P = 0.23). Fpn, ferroportin.
|
|
Figure 8.
Amino-acid sequence alignment of human and mouse ferroportin. We aligned the human Fpn (NCBI Reference Sequence: NP_055400.1) and mouse Fpn (NCBI Reference sequence: NP_058613.2) sequences by using the Basic Local Alignment Search Tool (BLAST). Transmembrane regions (TMs) were predicted as described (21) and shown by gray shading. Human and mouse ferroportin sequences share 90% identity (518 of 574 residues) at the amino-acid level and 93% similarity (536/574). Similar amino acids in the sequence alignment are indicated by + in the consensus sequence (gray text). The sequence alignment resulted in 1% gaps (7/574), indicated by –. Positions at which either protein contains a histidine residue are highlighted as follows: yellow, intracellular; green, transmembrane; or cyan, extracellular. ECL, extracellular loop; Fpn, ferroportin; ICL, intracellular loop.
|