Copper(II) binding properties of hepcidin
- 1.6k Downloads
Hepcidin is a peptide hormone that regulates the homeostasis of iron metabolism. The N-terminal domain of hepcidin is conserved amongst a range of species and is capable of binding CuII and NiII through the amino terminal copper–nickel binding motif (ATCUN). It has been suggested that the binding of copper to hepcidin may have biological relevance. In this study we have investigated the binding of CuII with model peptides containing the ATCUN motif, fluorescently labelled hepcidin and hepcidin using MALDI-TOF mass spectrometry. As with albumin, it was found that tetrapeptide models of hepcidin possessed a higher affinity for CuII than that of native hepcidin. The log K 1 value of hepcidin for CuII was determined as 7.7. CuII binds to albumin more tightly than hepcidin (log K 1 = 12) and in view of the serum concentration difference of albumin and hepcidin, the bulk of kinetically labile CuII present in blood will be bound to albumin. It is estimated that the concentration of CuII-hepcidin will be less than one femtomolar in normal serum and thus the binding of copper to hepcidin is unlikely to play a role in iron homeostasis. As with albumin, small tri and tetra peptides are poor models for the metal binding properties of hepcidin.
KeywordsPeptide Mass spectrometry Homeostasis Binding affinity Biomedicine
Alignment and conservation of some hepcidin sequences
Interestingly, the N-terminal region of human hepcidin is capable of binding CuII and NiII ions via an ATCUN (amino-terminal Cu-Ni)-binding motif comprising the three N-terminal residues (H2N-XXH) [7, 8]. In principle almost all hepcidin molecules can chelate CuII at the N-terminal region, as the majority possess the ATCUN motif, the exceptions amongst the most common sequences, being rat and mouse where position 3 is occupied by asparagine; tialpia, where position 3 is arginine and bass where, although position 3 is serine and position 1 is histidine. Thus even those sequences lacking the ATCUN motif are predicted to bind CuII. Several research groups have demonstrated the binding of transition metals to hepcidin, for instance Farnaud et al.  reported the binding of FeII and Balesaria et al.  reported that the regulation of hepcidin levels is influenced not only by FeII, but also ZnII, CdII and CuII. In view of this interest in the divalent metal binding properties of hepcidin, particularly the binding of CuII, we decided to investigate this interaction and to quantify the affinity constant of hepcidin for CuII, such that we could compare the affinity constant with other endogenous CuII binding proteins and in particular albumin.
In principle the structure of the CuII complex of human hepcidin could adopt a similar structure, as previously suggested by Tselepis et al.  (Fig. 1b) who reported the dissociation constant between hepcidin and CuII as ≪1 µmol/L . As CuII is a kinetically labile cation , both human albumin and human hepcidin might be expected to compete with each other. In view of this possibility it is important to establish an accurate affinity constant for the binding of CuII to hepcidin. For this reason we have monitored the binding of CuII to human hepcidin, fluorescent human hepcidin and three peptide models of the N-terminus of human hepcidin, in order to gain a better insight into the possible significance of this potentially important CuII-hepcidin interaction.
Materials and methods
All solutions were prepared with Milli-Q (18 MQ) water. The Fmoc amino acid derivatives were obtained from Bachem AG (4416 Bubendorf, Switzerland). DMF, DMSO, 3-(N-morpholino) propanesulfonic acid (MOPS), copper atomic standard solution, acetonitrile, nitrilotriacetic acid (NTA), potassium chloride, Cu(II) sulphate pentahydrate and other reagents were purchased from Sigma-Aldrich and used without further purification. Analytical grade volumetric hydrochloric acid (0.2092 mol/L) and HPLC grade water (Fisher) were used in the preparation of all solutions used for potentiometry.
In the hepcidin sequence (DTHFPICIFCCGCCHRSKCGMCCKT) Methionine 21 was replaced with a Lysine (M21K) which was protected on the ε-amino group with the (4,4-dimethyl-2,6-dioxocyclohex-1-ylidene)ethyl group (Dde). Peptide synthesis was carried out on a CEM Liberty 1 microwave peptide synthesizer with Fmoc-Thr(tBu)-Trt-Resin on a 0.2 mmol scale. All deprotection reactions were carried out with 20 % piperidine in dimethylformamide (v/v). Acylation reactions were carried out using 2-(6-chloro-1-H-benzotriazole-1-yl)1,1,3,3-tetramethylaminiumhexafluorophosphate (HCTU) (0.78 mmol) in the presence of 2,4,6-collidine (1.6 mmol) for 10 min using 25 W microwave power. Cysteine and histidine were activated with diisopropylcarbodiimide (0.8 mmol) and ethyl cyano(hydroxylimino) acetate (0.8 mmol). Boc-Asp(OtBu)-OH was incorporated at the N-terminus . Peptides DTH, GGHG and DTHFPIAIF-NH2 were prepared similarly on PEG-PS resin.
For the preparation of the labelled hepcidin, the peptide resin (400 mg, 69 μmol) was treated with 15 mL of hydrazine hydrate in dimethylformamide (2 %, v/v) to selectively remove the Dde group. The free amino group was acylated with either dichloro-6-carboxyfluorescein (65 mg, 173 μmol) or 6-carboxy tetramethylrhodamine (45 mg, 104 μmol), diisopropylcarbodiimide (27 μL, 173 μmol) and ethyl cyano(hydroxyimino) acetate (25 mg, 173 μmol) in 2 mL of DMF over 18 h.
The peptide resins were cleaved from the solid support and simultaneously deprotected using trifluoroacetic acid, phenol, water, thioanisole and 3,6-dioxo-1,8-octanedithiol (82.5/5/5/0.5/2.5) for 4 h and precipitated in diethyl ether and then freeze dried.
The linear labelled hepcidin was then dissolved in reducing buffer (100 mL) composed of tris hydrochloride (0.5 mol/L), EDTA (0.1 mmol/L) and guanidine hydrochloride (6 mol/L) at pH 8.0 and reduced with excess dithiothreitol (53 mg, 345 μmol). The fluorescent peptide solution was acidified to pH 2.5 and then purified by preparative HPLC on a Vydac 218TP54 peptide and protein column (2.5 × 30 cm). The folding buffer (6.8 L) was composed of ammonium bicarbonate (100 mmol/L), EDTA (0.1 mmol/L), guanidine hydrochloride (2 mol/L), reduced glutathione (2.10 g, 6.9 mmol) and oxidized glutathione (0.42 g, 0.69 mmol). The purified reduced peptide was dissolved in folding buffer (100 ml) containing 6 mol/L GdmHCl and added to the bulk of the folding buffer. The solution was stirred for 48 h and the solution was purified by preparative HPLC to produce the crude folded peptide.
All peptides were purified by preparative HPLC on a Vydac peptide and protein column 218 TP54 semipreparative column (1.0 × 30 cm) and the resulting fractions were analysed on a Jupiter HPLC column (5 µm, C18, 300 Å, 2 mm × 150 mm, Phenomenex) using a 1100 HPLC system (Agilent Technologies). Mass spectrometry analysis of the peptides was carried out using matrix-assisted laser desorption ionization on a Bruker Autoflex . An Nd:YAG laser (355 nm) was used to irradiate the sample. The instrument was calibrated with a range of peptides covering the mass range (500–4000 Da). Samples were prepared on an Anchor Chip ® with a saturated solution of α-cyano-4-hydroxy-cinnamic acid as the matrix in a mixture of acetonitrile and 0.1 % trifluoroacetic acid. Direct infusion experiments were carried out on a linear Ion Trap (Thermo). (S1,S8: S3,S6: S2,S4: S5,S7-tetracyclo-DTHFPICIFCCGCCHRSKCGMCCKT (Hepcidin), S1,S8: S3,S6: S2,S4: S5,S7-tetracyclo-DTHFPICIFCCGCCHRSKCGK(5-tetramethylrhodamine)CCKT (M12K-TMR-hepcidin) and S1,S8: S3,S6: S2,S4; S5,S7-tetracyclo-DTHFPICIFCCGCCHRSKCGK(Dichloro-5-carboxyfluoresceine)CCKT (M21K-dichloro-CF-Hepcidin) were successfully prepared. Both M21K-TMR-hepcidin and M21K-dichloro-CF hepcidin exhibited measurable potency when tested in cultured Madin-Darby canine kidney (MDCK) and T47D cell expressing ferroportin .
The potentiometric titration system used in the study of GGHG, GGH and DTH-NH2 and their CuII complexes comprised of an autoburette (Metrohm Dosimat 765 litre mL syringe) and Mettler Toledo MP230 pH meter with Metrohm pH electrode (6.0133.100) and a reference electrode (6.0733.100). A potassium chloride electrolyte solution (100 mmol/L) was used to maintain the ionic strength. The temperature of the test solutions was maintained in a thermostatic jacketed titration vessel at 25 ± 0.1 °C by using a Techne TE-8 J temperature controller. The solution under investigation was stirred vigorously during the experiment. All instruments were interfaced to a computer and controlled by a Visual Basic program. An automated titration adopted the following strategy: the pH of a solution was increased by 0.1 pH unit by the addition of potassium hydroxide solution (100 mmol/L) from the autoburette; when pH readings varied by <0.01 pH unit over a 8 s period, they were judged to be stable and an incubation period was activated. For pKa determinations, an incubation period of 1.5 min was adopted; for metal stability constant determinations, an incubation period of 3 min was adopted. The cycle was repeated automatically until the defined end point pH value was achieved. All the titration data were analysed with the pHab program . The species plot was calculated with the HYSS program [17, 18]. Analytical grade reagent materials were used in the preparation of all solutions.
Following the same procedure as described in 2.3, the spectrum of the solution was additionally recorded at the end of the incubation period. The test solution was circulated through a Hellem quartz flow cuvette by a Gilson Mini-plus #3 pump with speed capability (20 mL/min). The flow cuvette was mounted on an HP 8453 UV–visible spectrophotometer. For pKa determinations, a cuvette path length of 10 mm was used while for metal stability constant determinations, a cuvette path length of 50 mm was used.
Fluorescence spectra were obtained using a Perkin Elmer LS50B fluorometer. Excitation and emission frequencies of 552 and 576 nm were adopted, and data were recorded at room temperature. Briefly, sample volumes of 2 ml, with peptide concentrations of 0.5 µmol/L M21K-TMR-hepcidin in 0.25 mol/L MOPS buffer (pH 7.45)/acetonitrile (v/v = 50:50) and aqueous 2 µmol/L CuIISO4 solution (50 µl) was added every 5 min.
The peptides were dissolved in a standard copper solution to form a 1:1 peptide to copper preparation. The solution was serially diluted with 20 mmol/L ammonium bicarbonate to prepare peptide—CuII solutions of different concentrations.
For the preparation of the liquid matrix for MALDI-MS, 10 mg of α-cyano-4-hydroxycinnamic acid, 30 μL of diisopropylethylamine and 50 μL of 10 mmol/L ammonium phosphate in aqueous methanol (50 %) were vortexed and sonicated to obtain a solution. This solution was then diluted 30-fold with 20 mmol/L ammonium bicarbonate: methanol (1:1, v/v). The sample and the matrix (2 μL each) were mixed and then spotted onto a MALDI plate and allowed to evaporate to yield a liquid film. The samples were then analysed by MALDI-TOF.
In order to gain a better understanding of CuII binding by the ACTUN motif of hepcidin we have studied CuII binding to various hepcidin N-terminal peptides. Some are tri- and tetrapeptides and are limited to the ACTUN motif and one DTHFPIAIF-NH2, is a nonapeptide representing the N-terminal section of hepcidin with the cysteine being replaced by alanine. We have also studied CuII binding to human hepcidin and two fluorescently labelled human hepcidin analogues.
Synthesis of the peptides
The synthesis of the peptides proceeded without any major difficulties the percentage yields of the crude peptides GGHG, DTH-NH2 and DTHFPIAIF-NH2 were 20, 40 and 60 % respectively. All peptides were greater than 95 % pure as demonstrated by HPLC and corresponded to the correct structures as confirmed by electrospray ionization mass spectrometry (ESI–MS) and/or MALDI-TOF MS data. For the labelled peptides, the linear peptides were successfully synthesised on the solid support by replacing methionine21 with Lys(Dde), the Dde group was removed with hydrazine and the free amino group was labelled with either 6-carboxytetramethylrhodamine or 6-carboxy dichloro-carboxyfluoresceine. The peptide was cleaved from the solid support, purified and then folded under oxidative conditions containing oxidised and reduced glutathione to produce the products, which were characterised by HPLC and MALDI-TOF MS (see supplementary data). Both M21K-TMR-hepcidin and M21K-dichloro-CF hepcidin exhibited measurable potency when tested in cultured Madin-Darby canine kidney (MDCK) and T47D cell expressing ferroportin .
Potentiometric determination of the pKa values of GGHG and DTH-NH2
Stability constants of short peptides
pCu 7.4 II *
The predominant form at pH 7.4 is AH where the terminal amino group function remains protonated and the peptide lacks a net charge. The pH titration and resulting speciation plot for DTH-NH2 are presented in Fig. 2c. The pKa values are 7.9 (terminal amino group), 6.3 (imidazole group) and 2.8 (carboxy function) (Table 2). The predominant form at pH 7.4 is AH where the terminal amino group function remains protonated and the peptide lacks a net charge.
Spectrophotometric determination of CuII affinity constants for GGHG and DTH-NH2
These values are close to the reported CuII log conditional affinity constants for some analogous peptides, that correspond to the N-terminal sequence of human albumin (DAH), for instance DAHK-NH2 (13.8)  and DAH-NH2 (13.7) . It is significant that these two values are higher than the log conditional affinity constant for CuII and HSA .
In this investigation, we adopted spectrophotometric titration over the pH range 2.5–10.5, whereas in previous studies, data of the potentiometric titration over the pH range 2.0–7.8 was adopted [22, 23]. With both GGH and GGHG we observed a transition over the pH range 6.5–8.5 which was possibly not recorded by the previous workers. We attribute this transition to the formation of copper complex dimers as previously suggested by Aiba et al. .
Determination of the copper(II) affinity constants for fluorescently labelled hepcidin
The difference in this behaviour is likely to be related to the relative ease of the formation of a square planar CuII structure, involving the dissociation of two amide protons. Small restrictions in the geometry could lead to relatively large changes in the affinity constant. By analogy with albumin, it is likely that the higher value of Log K 1 (CuII) for the N-terminal peptides, when compared with fluorescently labelled hepcidins, results from an interaction of the ATCUN motif with the rest of the hepcidin structure. Indeed the C-terminal region of hepcidin is close to the N-terminus region in the native structure it being held in such a position by the numerous disulphide bonds present in the structure .
Determination of hepcidin copper(II) affinity constants by mass spectroscopy
Although the two fluorescently labelled hepcidin molecules are likely to interact with CuII in a similar fashion to that of the native molecule, the possibility exists that the fluorescent probe could interfere with the conformation of the N-terminus region. Consequently we studied the binding of CuII to a N-terminal model peptide of hepcidin (DTHFPIAIF-NH2) and human hepcidin by MALDI-TOF. This is a useful technique for the analysis of biological molecules but has two limitations. Firstly the crystallisation of the matrix and the sample on the plates results in a non homogeneous surface which leads to loss of resolution and secondly most sample preparations are carried out in an acidic environment where metal and/or protein complexes tend to dissociate. In view of these limitations we have used a liquid matrix approach  in order to perform the study at physiological pH values and to maintain the homogeneity of the sample. Furthermore as signal suppression is common in MALDI analysis, we omitted the use of an internal standard and the first measurement was carried out using the peptide with no added copper.
In summary log CuII affinity constants for the two fluorescently labelled hepcidins were 7.5 and 7.2 and the log CuII affinity constant for native hepcidin determined by mass spectrometry was 7.7–7.8. These values suggest that the presence of the fluorescent probe only slightly reduces the affinity for CuII. Significantly the log CuII affinity constant for labelled N-terminal nonapeptide bound to CuII is approximately tenfold tighter than the hepcidin molecule, again confirming the difficulty of using short peptides to model cations binding to polypeptides.
We conclude that native hepcidin binds CuII with a log K 1 value of 7.7–7.8. This value agrees with the previous estimate of Tselepis et al.  (Log K 1 ≫ 6) and is now sufficiently accurate to undertake speciation studies with potential endogenous CuII chelators, for instance albumin.
Hepcidin binds CuII with relatively high affinity (Log K 1 = 7.7), but is this value sufficiently high to compete with intracellular copper chaperone proteins and extracellular proteins, for instance albumin? The intracellular labile copper pool is vanishingly small (<femtomolar)  due to the necessity of preventing copper from competing with iron and manganese for enzyme active sites . At such low CuII concentrations, hepcidin will not be able to bind copper. In contrast the extracellular concentration of labile CuII is much higher, typically 1–2 μmol/L , and the majority of this pool is bound to albumin. The affinity of human serum albumin for CuII is logK 1 = 12  which is approximately 100,000 times higher than the value for hepcidin as reported in this study. As CuII is kinetically labile , it is expected to rapidly distribute between albumin and hepcidin. The serum albumin concentration is approximately 600 µmol/L, whereas that of hepcidin is of the order of 10 nmol/L. Thus the bulk of kinetically labile CuII, will be bound to albumin. It is estimated that the concentration of hepcidin coordinated to CuII in normal serum is <1 fmol/L. Thus under both intracellular and extracellular environments, hepcidin will be unable to compete with other endogenous CuII binding ligands and thus it is unlikely that the binding of copper to hepcidin plays a role in iron homeostasis.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.