Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: J Inorg Biochem. 2020 Feb 15;206:111040. doi: 10.1016/j.jinorgbio.2020.111040

Significance of conformation changes during the binding and release of chromium(III) from human serum transferrin

Kyle C Edwards 1, Hannah Kim 1, Riley Ferguson 1, Molly M Lockart 1, John B Vincent 1,*
PMCID: PMC7108967  NIHMSID: NIHMS1568808  PMID: 32088595

Abstract

Trivalent chromium has been proposed to be transported in vivo from the bloodstream to the tissues via endocytosis by transferrin (Tf), the major iron transport protein in the blood. While both Cr(III) binding and release from Tf have been proposed to be too slow to be physiologically relevant, recent kinetic studies under physiological conditions demonstrate that Cr(III) binding and release are sufficiently fast to occur during the time of the endocytosis cycle (circa 15 minutes). Consequently, the release of Cr(III) from human and bovine serum Tf has been examined under conditions mimicking an endosome during endocytosis. These studies have also found that Cr(III)2-Tf can exist in multiple conformations giving rise to different spectroscopic properties and different rates of Cr(III) release. Time-dependent spectroscopic studies of the binding and release of Cr(III) from human serum Tf have been used to identify three different conformations of Cr(III)2-Tf. The conformation of Cr(III)2-Tf used in most previous studies forms too slowly to be physiologically relevant and slowly releases Cr(III) in endosomal pH range. The conformation formed between 5 minutes to 60 minutes after the addition of Cr(III) to apoTf at pH 7.4 in 25 mM bicarbonate resembles the conformation of Cr(III)2-Tf in its complex with Tf receptor (TfR) and loses Cr(III) rapidly at endosomal pH, although not as fast as the Tf-TfR complex. The significance of these conformations and the potential role of Tf in detoxification of Cr(III) are described.

Keywords: Chromium, Transferrin, Conformation, Kinetics, Transport

Graphical Abstract

The binding of Cr(III) to transferrin under conditions modeling the blood plasma is accompanied and followed by a series of conformational changes. These conformations lose Cr(III) upon acidification at different rates. The physiologically relevant conformation for Cr(III)2-transferrin to bind to transferrin receptor is identified.

graphic file with name nihms-1568808-f0001.jpg

1. Introduction

Iron as the ferric ion is transported from the bloodstream to the tissues of mammals by the protein transferrin, Tf. Tf is selective for Fe3+ in a biological environment as its two metal sites are adapted to bind ions with large charge-to-size ratios; however, Tf may also transport Cr(III) under physiological conditions [1]. Tf is comprised of two lobes with approximately 40% sequence homology and nearly superimposable three-dimensional structures [2]. Each lobe possesses a metal ion-binding site, and each metal ion binds concomitantly with a synergistic anion, usually (bi)carbonate. The metal-binding sites are comprised of four protein provided ligands: two tyrosine residues, an aspartate reside, and a histidine residue (Figure 1). The binding and release of Fe(III) is accompanied by significant conformational changes in the Tf molecule. The apoprotein possesses a more open conformation; whereas in the “closed” Fe-loaded conformation, Tf binds to transferrin receptor (TfR). After binding to the receptor, Tf is brought into the cell by endocytosis where acidification of the endosome releases the Fe(III). Fusion of the endosome with the cell membrane releases and recycles the apotransferrin (apoTf). In humans, Tf is normally about 30% saturated with Fe, allowing it to potentially bind and transport other metal ions.

Figure 1.

Figure 1.

Open in a new tab

Proposed Cr(III) ligation in the C-terminal lobe metal-binding site of Tf. Numbering follows the sequence of human serum Tf.

In vivo studies have suggested that Tf is the physiological carrier of Cr(III) from the bloodstream to the tissues [3-6], although this has recently been questioned [7]. The rate at which Cr(III) binds to and is released from Tf has attracted recent attention [7-13]. The binding of Cr(III) to Tf is first order in bicarbonate [12,14]; in the presence of 25 mM bicarbonate (the concentration in blood), the binding of Cr(III) to Tf is sufficiently rapid to be physiologically significant, reaching equilibrium in 15 minutes [14]. The two metal binding sites bind Cr(III) with significantly different binding constant [9]. Recently, the release of Cr(III) from Tf at endosomal conditions has been shown to be rapid for weak binding site but slow for the tight binding site [15]. However, when Tf binds to TfR, the rate of Cr(III) loss from the Tf-TfR complex becomes sufficiently rapid for Tf to serve as the physiological transporter of Cr(III) from the bloodstream to the tissues [15].

However, recent studies have suggested that the binding of Cr(III) to transferrins may be more complicated than previously anticipated. For example, the two Cr-binding sites of human serum Tf have be shown to be readily distinguished by EPR (frozen solutions at 77 K) [16]. These two EPR features are present for Tf that has been stored in the presence of Cr(III) at pH just above neutral for two weeks under ambient bicarbonate concentrations [16] or for 2 d in 25 mM bicarbonate [12]. At near neutral pH, Cr(III) binds to both sites on the protein. At pH 4.8 to 5.9, Cr(III) binds to only one site [17]. This tighter binding site possesses an EPR features at g = 5.08 and 5.66 [18] and corresponds to the N-terminus lobe binding site [16]. The weaker binding site Cr(III) gives rise to EPR features at g = 5.42 (and also a feature at g ~ 2) and corresponds to the C-terminus lobe binding site [16,17]. Yet, the length of time for these two features to appear have not been examined in detail. This is a concern as the times required for the features to appear and not change with time are considerably longer than the time required for Cr(III) binding (at least in 25 mM bicarbonate) [12,14]. Curiously, while the EPR features arising from Cr(III) bound to conalbumin are similar to those of human serum Tf [12], the features for lactoferrin at g ~ 5.15 and 5.62 correspond to the C-terminal lobe weaker metal binding site while that at g ~ 5.43 corresponds to the N-terminal stronger metal binding site, reversed from the binding site g values in the blood serum Tf [18]). Thus, the type of transferrin (serum vs. egg vs. milk) despite apparently having identical ligands bound to the Cr(III) can have different environments about the Cr(III) centers giving rise to the different EPR features observed. Heat-treated and glycated human serum Tf while binding 2 Cr(III) per protein have been reported to generate only an EPR feature at g = 5.41 after sitting for 1 d after addition of Cr(III); with time, this feature loses intensity with the appearance of features at g = 5.14 and 5.64 [13]. Hence, a conformation of Cr(III)2-Tf appears to exist that gives rise to only a single feature in the g ~ 5 region. Consequently, the existence of multiple conformations during the Cr(III) uptake and release processes, giving rise to distinguishable spectroscopic features, would not be surprising. Similarly, studies on the uptake and release of Fe(III) by Tf or the Tf-TfR complex have revealed that several different conformational changes are involved in the processes, although exactly what these conformations are is controversial [19-21].

Herein are reported time-dependent spectroscopic studies of the binding and release of Cr(III) from human serum Tf that have been used to identify multiple different conformations of Cr(III)2-Tf and their role in the binding and release of Cr(III).

2. Materials and methods

2.1. Materials

Iron-free human serum Tf was obtained from Aldrich (St. Louis, MO). Doubly deionized water was used throughout. All reagents were used as received unless otherwise noted. Cr(III) solutions were prepared by using Cr(III)Cl3·6H2O. ApoTf concentrations were determined by using the extinction coefficient (ε = 9.12 × 104 M−1 cm−1) at 280 nm [22].

2.2. Methods

Dichromic-Tf was prepared as previously described [12]. For binding studies, samples of apo-Tf were prepared immediately beforehand by dissolving the protein (0.5 mM) in a buffered solution (Hepes (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), 100mM, and HCO3−, 25mM), and the pH of the resulting solution was adjusted to 7.5. A stock Cr(III) solution was prepared using CrCl3·6H2O in the Hepes/bicarbonate buffer solution but without adjusting the pH to 7.5. An aliquot of the Cr(III) solution was added to the apo-Tf solution to achieve a Cr(III) concentration of 1 mM; addition of the CrCl3 solution lowers the desired pH to 7.4. Addition of Cr(III) in this manner to the transferrin solution prevents the formation of Cr(III)-carbonate precipitates. Aliquots of the Cr(III)2-Tf solution were taken at prescribed intervals following the addition of the Cr(III) solution, and the aliquots were rapidly frozen for later EPR analysis.

To model the acidification of the endosome that triggers release of metal ions from Tf, Cr(III)2-Tf in a buffered solution (Hepes, 100 mM) at pH 7.4 containing 25 mM bicarbonate at 37 °C was acidified by the addition of hydrochloric acid to pH 4.5 or 5.5. After time prescribed intervals, aliquots were removed and frozen for analysis by EPR spectroscopy. Similar studies were performed in the presence of chelating ligands that could potentially accelerate Cr(III) release. Titrations of HCl with Cr2-Tf were performed beforehand to determine the quantity of HCl to be added to achieve desired pH. All results except ESEEM (Electron Spin Echo Envelope Modulation) measurements are presented as the average of at least triplicate experiments. Error bars in figures represent standard deviation.

2.3. Instrumentation

Ultraviolet-visible spectra were obtained by using a Cary (Aligent, Santa Clara, CA) 500 or Beckman Coulter (Brea, CA) DU800 UV-visible spectrophotometer. Binding of Cr3+ to Tf was monitored at 245 nm. Solutions were continuously stirred at 37 °C using a 6 × 6 Peltier thermostatted multicell holder. Continuous wave (CW) EPR were measured on a Bruker (Billerica, MA) ELEXSYS E540 X-band spectrometer with an ER 4102 ST resonator. CW spectra were measured at 9.44 GHz with a microwave power of 21.1 mW by using a magnetic field modulation frequency of 100 kHz with an amplitude of 30 Gauss. Spectra were taken at liquid nitrogen temperatures with a quartz insertion dewar. Electron spin echo envelope modulation (ESEEM) spectra were measured using an ELEXSYS E680 EPR spectrometer (Bruker-Biospin, Billerica, MA) equipped with a Bruker Flexline ER 4118 CF cryostat. ESEEM measurements were made at 5 K both at 3375 G (337.5 mT) and at 2020 G (202.0 mT). Data were collected using a three pulse stimulated echo sequence, π/2−τ−π/2−T−π/2−τ+T−echo, where π/2 represents a 16 ns microwave pulse, and T and τ represent delays between the pulses. The pulse sequence was repeated at a rate of 1.25 kHz. The delay time τ was chosen to be 120 ns, a value that was optimal for both echo intensity and nuclear modulation. Spectra were processed in the Xepr software (Bruker Biospin, Billerica MD). Plots were made using Origin2018 (OriginLab, Northhampton MA).

2.4. Data analysis

Data analysis, calculation of averages and standard deviations, and fitting of curves to the appropriate equations was performed by using SigmaPlot 11 (SPSS, Inc., Chicago, IL). The iterative curve fitting algorithm of SigmaPlot 11 uses the Marquardt-Levenberg algorithm to find the parameters of the independent variables that provide the best fit between the data and the equation.

CW EPR spectra processing and simulations were performed using the EasySpin package in MATLAB (Mathworks, R2017b) [23]. Polynomial fitting of the spectral baseline corrections was performed using Xepr software of the ELEXSYS. The g-values, g-strain, and weight of the simulated spectrum were fit using the “pepper” function in EasySpin [23].

3. Results and discussion

3.1. Binding of Cr(III) to transferrin

The binding of Cr(III) and the subsequent conformational changes of the resulting Cr(III)2-Tf complex has been found to occur in three phases under physiologically relevant conditions, i.e. pH 7.4 and 25 mM bicarbonate. Each phase can be distinguished by a combination of UV and EPR spectroscopies.

The EPR spectra of human serum Tf and other forms of Tf have been described above. Human serum Cr(III)2-Tf has been described as pale blue in color with visible maxima at 440 and 635 nm [16]. The visible spectra are typical for Cr(III) centers in a pseudo-octahedral environment. Most electronic spectral studies have been performed with conalbumin. Titrating conalbumin with Cr(III) and following the absorbance changes at 435 and 610 nm results in a linear increase in absorption until 2 equivalents of Cr(III) are added, after which the absorbance increases linearly but with a decreased slope; this is consistent with Cr(III) binding in both metal binding sites after which the increase in intensity results from the increase of free Cr(III) in solution [12]. Unfortunately, the intensity of the visible features prevents their use in monitoring changes in metal-binding to Tf at reasonable Tf concentrations. In the ultraviolet region, changes in the intensity of bands at ~245 and 295 nm only increase up to 2 equivalents of Cr(III) per conalbumin molecule [12], consistent with tight binding of Cr(III) in both metal binding sites as the π – π* transitions in the two tyrosine ligands are altered. The binding of two equivalents of chromic ions to conalbumin at ambient bicarbonate concentration was shown by Tan and Woodworth to lead to Δε at 245 nm of 38.7 × 103 (±2.8 × 103) M−1cm−1 after allowing two days for equilibrium to be achieved [24].

Fortunately, electronic spectral studies with human serum Tf yield results similar to those utilizing conalbumin. For human serum Tf, the change in Δε at 245 nm upon the addition of Cr(III) is similar to that for conalbumin, although after 48 h the Δε does not reach ~ 38 × 103 M−1cm−1 [12]. Yet, titration studies that show tight binding of chromic ions, whether at ambient or 25 mM bicarbonate concentrations have initial slopes Δε/Cr of well under 19 × 103 M−1cm−1 for both Tf and conalbumin. Similarly, the curves for binding of chromic ions to Tf and conalbumin as a function of time over 2 h and a function of bicarbonate concentration do not reach a value of Δε of ~ 38 × 103 M−1cm−1 [12].

These results indicate that tight binding of chromic ions to transferrins is accompanied by a change in extinction coefficient less than ~38 × 103 M−1cm−1, followed by other slower changes resulting in an increase in extinction coefficient to the final value. If using the assumption of Tan and Woodworth [24] that deprotonation of a tyrosine residue is accompanied by Δε245 of ~ 10 × 103 M−1cm−1, then the different rates of change in the extinction coefficient could suggest binding of Cr(III) is initially accompanied by binding to tyrosinate followed by a slower binding to a second tyrosinate.

3.1.1. Initial phase Formation of conformer 1

The initial phase involves the initial binding of Cr(III) and is complete in circa 15 minutes. The binding of Cr(III) is accompanied by the largest change in the extinction coefficient at 245 nm of the three phases, resulting from the binding of Cr(III) to tyrosine residues. This phase has previously been studied in detail [12]. The binding of Cr(III) to Tf is first order in bicarbonate [12,14] and occurs in the presence of 25 mM bicarbonate (the bicarbonate concentration in blood) with pseudo first order rate constants of 3.16 ± 0.13 and 0.189 ± 0.003 min−1 for binding of Cr(III) to the two metal binding sites [12]. The binding of two Cr(III) is also accompanied by an intense increase in the ultraviolet extinction coefficient at 245 nm of ~24 × 103 cm−1M−1.

The addition of CrCl3 to apoTf results in the conversion of the EPR feature from CrCl3 and its hydrolysis products (See Figures S1 and S2) into a symmetric feature at g ~ 2 (Figure 2). This feature arises from a short-lived intermediate/conformation, conformer 1, that reaches its maximum approximately 10 min after addition of Cr(III) to the Tf; the rate of appearance of this feature matches the average rate of Cr(III) binding to Tf in the presence of bicarbonate measured previously by ultraviolet spectroscopy. The binding of Cr(III) to Tf is stoichiometric in the time regime [12]. Thus, after ~ 10 min, the majority of Tf appears to be in the form of this transient, 1, species with two bound Cr(III) and with an EPR feature at g ~ 2.

Figure 2.

Figure 2.

Open in a new tab

EPR spectrum of Cr(III)2-Tf 5 min after addition of Cr(III) to apoTf in 100 mM HEPES with 25mM HCO3−, pH 7.4, at 37 °C. The major feature at g ~ 2 (~ 3,500 G) corresponds primarily to conformer 1 of Cr(III)2-Tf. A trace of the feature from conformer 2 at g ~ 5.4 (~ 1,200 G) can also be observed.

A crystal structure of Fe-loaded Tf in which the lobes are in the open confirmation has been reported [25]. Fe(III) is bound to both tyrosine residues in the partially formed metal-binding site. However, the lobes have not yet folded to the closed conformation that would bring the aspartate and histidine residues that bind Fe(III) in the closed conformation into close proximity of the Fe(III). This has led to the proposal that Fe binds first to the tyrosine residues followed by the closing of the metal-binding site accompanied by binding of the aspartate and histidine residues. If this were the case with Cr(III), then the initial intermediate formed upon Cr(III) binding should lack nitrogen coordination. However, ESEEM studies have revealed that this is not the case (Figure 3). Coupling to 14N, presumably of the imidazole ring of the histidine ligand, is readily detected. The binding of the histidine residue to Cr(III) not only indicates that the Tf is in the closed conformation but that two tyrosinates should be bound to the chromic centers. (The X-ray structure of Cr(III)-Tf reveals that Cr(III) binds to aspartate, histidine, and two tyrosinate residues along with two oxygens of the synerginic ligand in a similar fashion to Fe(III) [26]). Thus, the changes in extinction coefficient cannot be taken as evidence for the binding of only a single tyrosinate in this phase. From spectroscopic evidence, the Cr(III) centers of conformer 1 appears to have two tyrosine ligands and one histidine ligand provided the protein, while kinetic evidence suggests the involvement of a (bi)carbonate anion. Given that this places Cr(III) in the metal-binding site of each lobe, an aspartate ligand would also be expected, as well, to complete six-coordination about the Cr(III) centers.

Figure 3.

Figure 3.

Open in a new tab

Three-pulse ESEEM spectra of apo-Tf incubated with 2.0 equivalents of Cr(III) at 5, 60, and 1440 min following the addition of Cr(III). a - Feature at 4 MHz corresponds to coupling between 14N in the vicinity of the Cr(III). b - Feature at ~15 MHz corresponds to coupling between Cr(III) and nearby 1H’s. The intensities of these spectral features depend on the tau used during measurement.

3.1.2. Second phase Formation of conformer 2

The EPR feature from conformer 1 of Cr(III)2-Tf at g ~ 2 disappears concomitantly with the rise of an EPR feature at g = 5.42 (Figure 4) that reaches a maximum ~ 120 min after the addition of Cr(III) (Figure 5). The appearance of this EPR feature is accompanied by an increase in the extinction coefficient in the ultraviolet absorption of the Tf at 245 nm by ~ 8 × 103 M−1cm−1 (Figure 6). Thus, this phase represents a conformational change in the protein structure that involves the conformation of at least one tyrosine residue resulting in a change in the Cr(III) environment, with the product being conformer 2. Note that these conformation changes are being followed by the ultraviolet spectrum of the coordinated tyrosine residues and the EPR spectrum of the Cr(III) centers. Thus, the spectroscopy can only observe changes in the environment of the Cr(III) centers and their ligands. These changes require some movement of the protein to accommodate local to the metal-binding sites; however, global conformational changes, particularly of the magnitude of the change from open to closed conformations, are unlikely.

Figure 4.

Figure 4.

Open in a new tab

EPR spectra at time intervals after the addition of Cr(III) to apoTf in 100 mM HEPES buffer with 25 mM HCO3−, pH 7.4, at 37 °C.

Figure 5.

Figure 5.

Open in a new tab

Changes in the amplitude of the EPR features from conformer 2 (solid circles and line) and conformer 3 (open circles and dashed line) of Cr(III)2-Tf as a function of time.

Figure 6.

Figure 6.

Open in a new tab

Change in extinction coefficient at 245 nm as a function of time corresponding to the formation of conformations of Cr(III)2-Tf following the addition of Cr(III) to apoTf in 100 mM HEPES with 25 mM HCO3−, pH 7.4, at 37 °C. Inset: EPR spectra of aliquots were taken at prescribed intervals concurrent with the UV measurements.

Previously, the EPR feature at g ~ 5.42 has been assigned to Cr(III) in exclusively the weaker C-terminal metal binding site of Tf. However, this feature represents Cr(III) bound to both metal binding sites, as this feature decays with time to yield features corresponding to conformer 3 (vide infra) of Cr(III)2-Tf with features from both the C-terminal and N-terminal metal binding sites. Thus, conformer 2 is characterized by binding two Cr(III) ions while possessing a EPR feature at g ~ 5.42 (but not at g ~ 5.1 and 5.6). Cr(III) in both metal binding sites of conformer 2 giving rise to an EPR feature at g ~ 5.42 helps to explain some previously published results. Most notably, heat-treated and glycated human serum Tf while binding 2 Cr(III) per protein have been reported to generate only an EPR feature at g = 5.41 after sitting for 1 d after addition of Cr(III); with time, this feature loses intensity with the appearance of features at g = 5.14 and 5.64 [13]. Bovine Cr(III)2-Tf prepared by maintaining Tf in the presence of Cr(III) for 2 d in 25 mM bicarbonate also gives rise to a single feature at g ~ 5.4 unlike the human version with the three features at g ~ 5.08, 5.42, and 5.66 [15].

ESEEM studies confirm that N is coordinated to the Cr(III) in conformer 2 of Cr(III)2-Tf (Figure 3), again indicating the closed confirmation of the lobes of Tf. Consequently, the environment surrounding the Cr(III) centers in both metal binding sites must be very similar in conformer 2 to that in conformer 1, without a change in ligands to the Cr(III) centers.

3.1.3. Third phase Formation of conformer 3

After the EPR feature at g ~ 5.42 from conformer 2 reaches a maximum in intensity, the feature decays with the concomitant appearance of EPR features at g ~ 5.66 and 5.08 (Figures 4 and 5) corresponding to the formation of conformer 3. With time, the EPR spectrum resembles a spectrum previously associated with Cr(III)2-Tf; this is the feature for Cr(III)2-Tf observed after storage at ambient bicarbonate concentrations for two weeks or in 25 mM bicarbonate for 2 d. Thus, this spectrum appears to correspond to yet another conformation of Cr(III)2-Tf. The decay of the g ~ 5.42 feature and rise of the g ~ 5.66 and 5.08 features is accompanied by a decrease in the ultraviolet extinction coefficient at 245 nm, suggesting another conformational change in at least the N-terminal lobe involving a tyrosine ligand. Conformer 3 is, thus, distinct from conformers 1 and 2 by possessing EPR features at g ~ 5.08, 5.42, and 5.66.

Generating Cr(III)2-Tf in buffer with ambient bicarbonate over two weeks yields a form with an identical EPR spectrum to the one for Cr(III)2-Tf generated during this third phase, i.e. conformer 3. However, the ultraviolet spectra of these two forms of Cr(III)2-Tf are not identical; the species generated at ambient carbonate concentration has an extinction coefficient at 245 nm of 38.7 × 103 (± 2.8 × 103) M−1cm−1 [24], compared to that of ~24 × 103 M−1cm−1 for conformer 3 about 1 day after addition of Cr(III) to apoTf in 25 mM bicarbonate.

ESEEM studies confirm that N is coordinated to the Cr(III) in conformer 3 (Figure 3). Consequently, the environment surrounding the Cr(III) centers in both metal binding sites must be very similar in conformer 3 to those in conformers 1 and 2, without a change in ligands to the Cr(III) centers.

3.1.4. Summary of Cr(III)2-Tf species

Overall, the formation of three conformers of Cr(III)2-Tf after the addition of Cr(III) to apoTf can be observed by EPR and ultraviolet spectroscopy. While the rate constant for the formation of conformer 1 can be determined using ultraviolet spectroscopy, the rate of formation of the subsequent two conformers can only be estimated. This is limited, for example, by the inability to determine the concentration of the first conformer (with the g ~ 2 EPR feature) by EPR or UV spectroscopy. This broad feature at g ~ 2 from this species overlaps the broad EPR features from CrCl3 (and its hydrolysis products), the g ~ 2 component of the EPR feature from conformer 2, and the g ~ 2 component of the EPR feature of the C-terminal Cr(III) from conformer 3. Similarly, the relative areas of the EPR features in the region g ~ 5 to 6 allow for determination of the relative amounts of the conformers 2 and 3. However, the distribution of the conformations as a function of time could be simulated using the integrated rate equations for three consecutive irreversible reactions [27] with k1 = 2.4, k2 = 4.3 × 10−2, and k3 = 1.5 × 10−3 min−1 (Figure 7).

Figure 7.

Figure 7.

Open in a new tab

Simulated distribution with time of apoTf and conformations of Cr(III)2-Tf species following the addition of Cr(III) to apoTf. Solid line – apoTf; dotted line – conformer 1; dashed line – conformer 2; and dotted and dashed line – conformer 3.

3.1.5. Comparison with Fe(III) and Co(III) binding

Stopped-flow fluorescence studies of Fe(III) binding to Tf have identified three to four processes and four species involved in Fe(III) binding [19]. The first process involves iron-binding concomitantly with bicarbonate binding in the C-terminal lobe. This process is complete in about 0.1 s. This is followed by a second step involving two proton losses and the binding of the tyrosine phenolate ligands and requiring about 1 s for completion. This, in turn, is followed by a protein conformational change and proton loss, is complete in about 400 s, and triggers rapid Fe(III) incorporation into the N-terminal site followed by formation of the final conformation. The final conformation is achieved in the slowest process that is complete in about 3000 s [19]. Comparison with the results of the Cr(III)-binding studies suggests that the initial phase of Cr(III) binding detected by ultraviolet spectroscopy encompasses the first three steps in Fe(III) binding and the subsequent rapid binding of the second Fe(III). Thus, the time required for the binding of two Cr(III) to serum Tf is similar to that for the binding of the two Fe(III). For example, this would explain the first order dependence on bicarbonate of the initial phase of Cr(III) binding. The final step in Fe(III) binding in which the final equilibrium conformation is achieved appears to be separable into two phases for Cr(III) binding, both of which occur slower than the final step for Fe(III).

Comparisons with the binding of Co(III) to Tf can also prove useful. While the binding of Co(III) and carbonate from a pre-assembled Co(III)-carbonate complex to the C-terminal lobe is very rapid and requires only ~ 3 ms, this is followed by three slower processes lasting from 50 s for the binding of the two tyrosinates to 24 h [28]. The final two processes are accompanied by the binding of a second Co(III) in the N-terminal site, proton losses, and conformational changes. These slower processes start to resemble the time frames for the slower phases associated with Cr(III) binding; however, the steps could not be well resolved using a single technique for Co(III) binding.

3.2. Release of Cr(III) from transferrin

Recently, under conditions to model endocytosis, the loss of Cr(III) from Cr(III)2-Tf at pH 4.5 and 5.5 was shown to correspond to loss of Cr(III) rapidly from the weaker C-terminal metal binding site and much slower loss from the tighter N-terminal metal binding site [15]. The rate of release could be accelerated slightly by the addition of chelating ligands; however, the rates of Cr(III) loss were independent of the nature of the chelating ligand, suggesting the ligand might be assisting in loss of the synergistic (bi)carbonate ligand, rather than serving as a ligand to Cr(III). In fact ligand binding to Cr(III) was found to be too slow compared to the Cr(III) loss from Tf for the chelating ligands to be binding to Cr(III) [15]. When Cr(III)2-Tf was bound to the Tf receptor, Cr(III) loss from both binding sites was sufficiently rapid to occur during the endocytosis cycle [15].

Previously, in addition to the EPR features for Cr(III) in the two tight Cr(III)-binding sites on Tf disappearing upon acidification, new EPR features appear over time due to weak binding of Cr(III) to Tf [15]. Similar features are observed if Cr(III) is added to Tf at pH 5.5. However, the short time for Cr(III) to be lost from Tf compared to the long time required for the weak binding features to appear indicates that other forms of Cr(III) are present in the interim time period. This has been assumed to be free Cr(III). The result of the current study on the binding of Cr(III) raises the question of whether similar conformational changes occur during Cr(III) loss from Cr(III)2-Tf. These conformations certainly would be expected if Cr(III) loss was the reverse of Cr(III) binding. However, binding of Cr(III) at neutral pH is not the reverse of Cr(III) loss at acidic pH’s. For example, bicarbonate binding to Tf precedes [29] or potentially accompanies Cr(III) binding, whereas upon acidification bicarbonate loss almost certainly occurs before Cr(III) loss.

In the human circulatory system, the half-life of Tf is 7.6 days, while the lifetime of Tf-bound Fe is 1.7 h [30]. Given this lifetime for Cr(III) bound to Tf, the two conformers of Cr(III)2-Tf of potential biological relevance would appear to be the conformer 1 giving rise to the symmetric g ~ 2 EPR feature and conformer 2 with the EPR feature at g ~ 5.4, with conformer 3 giving rise to distinct features from the C-terminal and N-terminal binding sites forming too slowly to be relevant in vivo. However, conformer 3 has been the one traditionally studied as representative of Cr(III)2-Tf [1, 16-18]. Conformer 1 has a lifetime that is short enough to be of limited significance. The EPR feature from conformer 2 also has a nearly identical appearance to that of Cr(III)2-Tf bound to TfR [15], suggesting both forms could have a similar conformation. Thus, two new conformers of Cr(III)2-Tf have been observed in this work, conformers 1 and 2. Both may be of physiological significance based on the short lifetime of Fe(III) bound to Tf; however, the conformer of primary physiological significance is new conformer 2. The conclusions of previous investigations of Cr(III)-Tf should be carefully evaluated based on these considerations.

As the loss of Cr(III) from conformer 2 (whose lifetime is similar to that of Fe2-Tf in the bloodstream) has not been probed previously, this loss was examined. Loss of Cr(III) from Cr(III)2-Tf generated by storage at pH 7.4 with 25 mM bicarbonate for 1 d (i.e., representing almost completely the third conformation) by acidification to pH 5.5 can be followed readily by a combination of EPR and UV spectroscopies (Figure 8). Within the first 15 minutes, the C-terminal, weak binding Cr(III) is selectively lost with the concomitant loss of the EPR feature at g ~ 5.4 and a loss of half the UV absorption intensity at 245 nm. This is followed by a slower loss of Cr(III) from the tighter N-terminal binding site that leads to the loss of the EPR features at g ~ 5.1 and 5.6 and a continued loss of absorption intensity. With time EPR features between g = 5 and g = 6 appear from Cr(III) loosely associated with Tf.

Figure 8.

Figure 8.

Open in a new tab

Decrease of the extinction coefficient at 245nm of Cr2-Tf blanked against apo-Tf following lowering of pH to pH 5.5. Acidification initiated after 24 hr of Cr(III) incubation with Tf (2 Cr(III):1Tf) at pH 7.4, 37 °C (Fig 5). Inset: EPR spectra of aliquots were taken at prescribed intervals concurrent with the UV measurements.

Loss of Cr(III) from conformer 2 (allowed to form for 1 h after the addition of CrCl3 to apoTf) after acidification to pH 5.5 was also followed by EPR spectroscopy (Figure 9), although some conformer 3 is also necessarily present (See Figure 7). The g ~ 5.4 EPR feature (primarily from conformer 2 with some contribution from conformer 3) is lost rapidly, while the g ~ 5.2 and 5.6 features from conformer 3 are lost more slowly and four features from weakly bound Cr(III) appear over time. Loss of the g ~ 5.4 and the g~5.2 and 5.6 features (See Figure S3) can readily be fit to an exponential loss giving rise to first order rates constants of 0.113 and 0.0044 min−1, respectively, corresponding to half-lives of 5.21 and 1.6 × 102 min (Figure 10). These values are nearly identical to those for the loss of these features from Cr(III)2-Tf prepared by incubating Cr(III) and Tf for 2 d at 37 °C pH 7.4 in 25 mM bicarbonate [15]. Thus, the rate of loss upon acidification to pH 5.5 of both Cr(III) from conformer 2 is similar to the loss of the Cr(III) giving rise to the g~5.4 feature for Cr(III)2-Tf prepared in ambient bicarbonate concentrations (giving rise to the three EPR features at g ~ 5.2, 5.4, and 5.6). Hence, the conformation about a Cr(III) center giving rise to the g~5.4 EPR features allows for more ready loss of Cr(III) upon acidification than the conformation about a Cr(III) center giving rise to the g ~ 5.2 and 5.6 features, regardless of whether Cr(III) is in the C-terminal or N-terminal metal binding site. Neither loss is nearly as rapid as loss of Cr(III) from the Cr(III)2-Tf and TfR complex with a half-life of 0.242 min [15].

Figure 9.

Figure 9.

Open in a new tab

EPR spectra of Cr(III)2-Tf (formed by incubation of Cr(III) and Tf for 60 min) in 100 mM HEPES with 25 mM HCO3−, pH 7.4, at 37 °C at time intervals following the lowering of pH to 5.5.

Figure 10.

Figure 10.

Open in a new tab

Percentage loss of Cr(III) from Cr(III)2-Tf’s after acidification to pH 5.5. Black triangles - loss of Cr(III) from primarily conformer 2 (g ~ 5.4 feature) of Cr(III)2-Tf (formed 1 h after the addition of CrCl3 to apoTf in 100 mM HEPES with 25 mM HCO3−, pH 7.4, at 37 °C). Black circles - loss of the Cr(III) from the N-lobe site (g ~ 5.2 and 5.6 features) of conformer 3 (formed 48 h after the addition of CrCl3 to apoTf in 100 mM HEPES with 25 mM HCO3−, pH 7.4, at 37 °C).

The g ~ 2 feature of conformer 1 that is observed shortly after binding of Cr(III) to Tf was not observed at any point following the acidification of any conformation of Cr(III)2-Tf. This is expected given that the mechanism of Cr(III) loss at pH 5.5 is not the reverse of that of the binding of Cr(III) at pH 7.4. The loss of intensity of the UV feature at 245 nm upon acidification of Cr(III)2-Tf (Figure 8) reveals that loss of the EPR features associated with the bound Cr(III) is accompanied by loss of the binding of Cr(III) to the tyrosine residues of the metal binding site; thus, Cr(III) is being released from the metal binding sites. In the case of acidification, loss of the EPR features does not arise from interconversion of conformations but must result from loss of Cr(III) from Tf.

Given the short lifetime of conformer 1 (with the g ~ 2 EPR feature) and that significant amounts of free Cr(III) and/or conformer 2 (both of which also give rise to broad EPR features at g ~ 2) overlap, no attempts were made to study Cr(III) release from conformer 1.

3.3. Physiological significance

For Cr(III), the fate of injected 51Cr(III)2-Tf has been followed in rats [31, 32]. Greater than 50 % of labelled Cr(III) from Tf migrates to the tissues within 30 min of injection, and tissue levels of labelled Cr(III) are maximal 30 minutes after injection. [32]. Subsequently, 51Cr(III) moves from the tissues to the urine such that ~50 % of the 51Cr(III) is excreted in the urine within 6 h of injection [32]. The vast majority of the labelled Cr(III) in tissues is found in the liver and skeletal muscle. About 90% of injected 51Cr(III)2-Tf is lost in the bloodstream in a rapid process with a rate constant of ~0.084 min−1 (t1/2 ~ 8 min). 51Cr(III) in the hepatocytes is primarily associated with Tf as 51Cr(III) elutes in a band with a molecular weight corresponding to that of Tf with hepatocyte contents are applied to S-200 columns and immunoblotting of the contents of the band indicates the presence of Tf [31]. The rate of movement of 51Cr(III) from injected 51Cr(III)2-Tf to the tissues is accelerated by co-injection of insulin, as expected for a Tf-mediated process [31, 32]. When Tf is stored at elevated temperature for two weeks or glycated and then loaded with 51Cr(III), the amount of 51Cr(III) in the bloodstream 2 h after injection is greatly increased (particularly for the glycated Tf) while the amount of 51Cr(III) in the tissues is greatly reduced [13]. While the movement of 51Cr(III) from the bloodstream to the tissues is still enhanced by co-injection of insulin, the insulin only results in an increase in 51Cr(III) in the liver and not in the muscle whose 51Cr(III) levels are unchanged. This is in stark contrast to the behavior using freshly prepared 51Cr2-Tf where 51Cr(III) levels increase similarly in liver and muscle in response to insulin [32] and suggests faster catabolism of Tf in response to insulin rather than increased Tf-receptor mediated transport. Thus, Cr(III) appears to be transported from the bloodstream to the tissues by Cr2-Tf, and this occurs with a lifetime for the Tf-bound Cr that is probably not too different from that of Tf-bound Fe.

4. Conclusion

A combination of electronic and EPR studies reveal that the addition of Cr(III) to apoTf at near neutral pH in the presence of 25 mM bicarbonate results in the rapid binding of Cr(III) accompanied and then followed by a series of conformation changes in the protein-Cr(III) complex. These multiple conformations give rise to different spectroscopic properties and upon acidification different rates of Cr(III) release. Thus, knowledge of which conformation of Cr(III)2-Tf could be important in interpreting results of studies on Cr(III) release from Cr(III)2-Tf. The conformer of Cr(III)2-Tf used in most previous studies and giving rise to EPR features at g ~ 5.1, 5.4, and 5.6 forms too slowly to be physiologically relevant and slowly releases Cr(III) in endosomal pH range. Most significantly these studies identified two unknown conformers of Cr(III)2-Tf that give rise to an EPR feature a g ~2 and an EPR signal a g ~ 5.4, respectively. The latter of these conformers has a lifespan similar to the turnover time of transferrin and releases Cr(III) rapidly, suggesting it is probably the most physiologically significant conformer of Cr(III)2-Tf.

Supplementary Material

1

Highlights:

  • Binding of Cr(III) to transferrin involves a series of conformation changes

  • Conformations of Cr(III)2-transferrin can probed by UV and EPR spectroscopies

  • Conformation of Cr(III)-transferrin that binds transferrin receptor is identified

  • The conformations of Cr(III)-transferrin lose Cr(III) at different rates at pH 5.5

Acknowledgements

K.C. E. was supported in part by Department of Education GAANN Grant #P200A150329. H. K. was supported in part by an Undergraduate Research and Creativity Academy award from the College of Arts and Sciences of The University of Alabama. M.M.L. was supported by NIH-NIGMS (2 R15 GM117511-02).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  • [1].Vincent JB, Love S, Biochim Biophys. Acta 1820 (2012) 361–378. [DOI] [PubMed] [Google Scholar]
  • [2].Baker EN, Adv. Inorg. Chem 41 (1994) 389–463. [Google Scholar]
  • [3].Hopkins LL Jr., Schwarz K, Biochim. Biophys. Acta 90 (1964) 484–491. [DOI] [PubMed] [Google Scholar]
  • [4].Clodfelder BJ, Vincent JB, J. Biol. Inorg. Chem 10 (2005) 383–393. [DOI] [PubMed] [Google Scholar]
  • [5].Clodfelder BJ, Emamaullee J, Hepburn DDD, Chakov NE, Nettles HS, Vincent JB, J. Biol. Inorg. Chem 6 (2001) 608–617. [DOI] [PubMed] [Google Scholar]
  • [6].Clodfelder BJ, Upchurch RG, Vincent JB, J. Inorg. Biochem 98 (2004) 522–533. [DOI] [PubMed] [Google Scholar]
  • [7].Levina A, Nguyen Pham TH, Lay PA, Angew. Chem. Int. Ed 55, 2016, 1–5. [DOI] [PubMed] [Google Scholar]
  • [8].Quarles CD Jr., Marcus RM, Brumaghim JL, J. Biol. Inorg. Chem 16 (2011) 913–921. [DOI] [PubMed] [Google Scholar]
  • [9].Sun Y, Ramirez J, Woski SA, Vincent JB, J. Biol. Inorg. Chem 5 (2000) 129–136. [DOI] [PubMed] [Google Scholar]
  • [10].Quarles CD Jr., Brumaghim JL, Marcus RK, Metallomics 2 (2010) 792–799. [DOI] [PubMed] [Google Scholar]
  • [11].Li Y, Liu B, Yang B, J. Luminescence 130 (2010) 2339–2345. [Google Scholar]
  • [12].Deng G, Wu K, Cruce A, Bowman MK, Vincent JB, J. Inorg. Biochem 143 (2015) 48–55. [DOI] [PubMed] [Google Scholar]
  • [13].Deng G, Dyroff SL, Lockart M, Bowman MK, Vincent JB, J. Inorg. Biochem 164 (2016) 26–33. [DOI] [PubMed] [Google Scholar]
  • [14].Vincent JB, Edwards KC, in Vincent JB (Ed), The Nutritional Biochemistry of Chromium(III), 2nd ed, Elsevier, Amsterdam, 2019, pp. 129–174. [Google Scholar]
  • [15].Edwards KC, Kim H. Vincent JB, J. Inorg. Biochem, in press. [Google Scholar]
  • [16].Aisen P, Aasa R, Redfield AG, J. Biol. Chem 244 (1969) 4628–4633. [PubMed] [Google Scholar]
  • [17].Harris DC, Biochemistry 16 (1977) 560–564. [DOI] [PubMed] [Google Scholar]
  • [18].Ainscough EW, Brodie AM, Plowman JE, Bloor SJ, Sanders Loehr J, Loehr TM, Biochemistry 19 (1980) 4072–4079. [DOI] [PubMed] [Google Scholar]
  • [19].El Hage Chachine JM, Hemadi M, Ha-Doung N-T, Biochim. Biophys. Acta 1820 (2012) 334–347. [DOI] [PubMed] [Google Scholar]
  • [20].Terpstra T, McNally J, Han T-H-L, Ha-Doung N-T, El-Hage-Chahine J-M, Bou-Abdallah F, J. Inorg. Biochem 136 (2014) 24–32. [DOI] [PubMed] [Google Scholar]
  • [21].Steere AN, Byrne SL, Chasteen ND, Mason AB, Biochim. Biophys. Acta 1820 (2012) 326–333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Aramini JM, Vogels HJ, J. Am. Chem. Soc 115 (1993) 245–252. [Google Scholar]
  • [23].Stoll S, Schweiger A, J. Magn. Reson 178 (2006) 42–55. [DOI] [PubMed] [Google Scholar]
  • [24].Tan AT, Woodworth RC, Biochemistry 8 (1969) 3711–3716. [DOI] [PubMed] [Google Scholar]
  • [25].Mizutani K, Yamashita H, Kurokawa H, Mikami B, Horise M, J. Biol. Chem 274 (1999) 10190–10194. [DOI] [PubMed] [Google Scholar]
  • [26].Vincent J, Edwards K, Petersen C, Gilbert N, Thompson M, Curr. Dev. Nutr 3 (suppl. 1) (2019) 2053. [Google Scholar]
  • [27].Rutherford E, Chadwick J, Ellis CD, Radiations from Radioactive Substances. MacMillan Company, New York, 1930, pp. 10–18. [Google Scholar]
  • [28].Chikh Z, Hemadi M, Miquel G, Ha-Doung N-T, El Hage Chahine J-M, J. Mol. Biol 380 (2008) 900–916. [DOI] [PubMed] [Google Scholar]
  • [29].Harris WR, Biochem. Biophys. Acta 1820 (2012) 348–361. [DOI] [PubMed] [Google Scholar]
  • [30].Aisen P, Metal Ions Biol. Sys 35 (1998) 585–631. [PubMed] [Google Scholar]
  • [31].Clodfelder BJ, Upchurch RG, Vincent JB, J. Inorg. Biochem 98 (2004) 522–533. [DOI] [PubMed] [Google Scholar]
  • [32].Clodfelder BJ, Vincent JB, J. Biol. Inorg. Chem 10 (2005) 383–393. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1568808-supplement-1.docx (214.4KB, docx)

RESOURCES