Complexation of ytterbium to human transferrin and its uptake by K562 cells Xiu-lian Du 1 , Tian-lan Zhang 1 , Lan Yuan 1 , Yong-yuan Zhao 1 , Rong-chang Li 1 , Kui Wang 1 , Siu Cheong Yan 2 , Li Zhang 2 , Hongzhe Sun 2 and Zhong-ming Qian 3 1 Department of Chemical Biology, School of Pharmaceutical Sciences, Peking University, Beijing, China; 2 Department of Chemistry and Open Laboratory of Chemical Biology, University of Hong Kong, Hong Kong; 3 Department of Applied Biology and Chemical Technology, Hong Kong Polytechnic University, Kowloon, Hong Kong There is an increasing interest in the use of lanthanides in medicine. However, the mechanism of their accumulation in cells is not well understood. Lanthanide cations are similar to ferric ions with regard to transferrin binding, suggesting transferrin-receptor mediated transport is possible; however, this has not yet been confirmed. In order to clarify this mechanism, we investigated the binding of Yb 3+ to apo- transferrin by UV-Vis spectroscopy and stopped-flow spec- trophotometry, and found that Yb 3+ binds to apotransferrin at the specific iron sites in the presence of bicarbonate. The apparent binding constants of these sites showed that the affinity of Yb 3+ is lower than that of Fe 3+ and binding of Yb 3+ in the N-lobe is kinetically fav- ored while the C-lobe is thermodynamically favored. The first Yb 3+ bound to the C-lobe quantitatively with a Yb/ apotransferrin molar ratio of < 1, whereas the binding to the other site is weaker and approaches completeness by a higher molar ratio only. As demonstrated by 1 H NMR spectra, Yb 3+ binding disturbed the conformation of apotransferrin in a manner similar to Fe 3+ . Flow cytometric studies on the uptake of fluorescein isothiocyanate labeled Yb 3+ -bound transferrin species by K562 cells showed that they bind to the cell receptors. Laser scanning confocal microscopic studies with fluorescein isothiocyanate labeled Yb 3+ -bound trans- ferrin and propidium iodide labeled DNA and RNA in cells indicated that the Yb 3+ entered the cells. The Yb 3+ -trans- ferrin complex inhibited the uptake of the fluorescein labeled ferric-saturated transferrin (Fe 2 -transferrin) complex into K562 cells. The results demonstrate that the complex of Yb 3+ -transferrin complex was recognized by the transferrin receptor and that the transferrin-receptor-mediated mech- anism is a possible pathway for Yb 3+ accumulation in cells. Keywords: K562 cells; recognition; transferrin; ytterbium. Lanthanides have been suggested for the treatment of a series of diseases and for diagnosis by magnetic resonance imaging [1,2]. Recent studies also show that they could act as scavengers of free radicals [3] and therefore protect cells and tissues from oxidative stress-induced injury. Some lanthanides nuclides were also suggested for palliative therapy. 169 Yb (c-emission, t ½ % 32 days) was reported to provide comparable tumor control and has been considered as a potential replacement for 125 Iand 103 Pd in permanent implants [4,5]. Evidently, the intracellular accumulation is very important in these cases, but its mechanism still remains unclear. It was suggested that the particulate- and protein-bound Ln enters the cells by endocytosis [6]; the anionic low- molecular-mass complexes, via anion channels [7], whereas free Ln 3+ is transported by ionophores [8], Na + /Ca 2+ exchange [9] and self-facilitated diffusion [10]. It is known that Ln 3+ is mainly bound to proteins in the extracellular media (e.g. plasma). Various studies have demonstrated that considerable amounts of Ln are bound to the iron transport protein transferrin (Tf) in the blood [11–14]. As metal ions of therapeutic and diagnostic interest also bind to Tf at the specific iron sites [15], Tf has been thought of as a Ôdelivery vehicleÕ for metal ions into cells [16–19]. Tf takes up Fe 3+ at pH 7.4 and transports it into cells via receptor-mediated endocytosis. In this transport system, ferric ion binds to apotransferrin (apo-Tf) first to form an iron-loaded Tf, and subsequently the iron-loaded Tf binds to the specific Tf receptor (TfR). The Tf-TfR complex is internalized and iron dissociates from Tf upon acidification of cytoplasm (pH % 5.5). The molecular recognition between Tf and TfR is believed to be critical for the iron transport. The recognition depends on the conformation of the protein, which is regulated by the metal ion, from the Ôlobe-openÕ state in the apo-form to Ôlobe-closedÕ state in the holo-form as revealed from X-ray crystal structures of Tf and the recombinant N-lobe of Tf [20–22]. Although Yb 3+ hasbeenreportedtobindtoapo-Tf,it is not clear whether Yb 3+ can enter the cells by the same way. Correspondence to K. Wang, Department of Chemical Biology, School of Pharmaceutical Sciences, Peking University, Beijing, 100083, China. Fax: + 86 10 6201 5584, Tel.: + 86 10 6209 1539, E-mail: wangkui@bjmu.edu.cn or H.Z. Sun, Department of Chemistry and Open Laboratory of Chemical Biology, University of Hong Kong, Pokfulam Road, Hong Kong. Fax: + 852 2857 1586, Tel.: + 852 2859 8974, E-mail: hsun@hkucc.hku.hk Abbreviations: apo-Tf, apotransferrin; Fe 2 -Tf, ferric-saturated trans- ferrin or holotransferrin; FITC, fluorescein isothiocyanate; FITC- Fe 2 -Tf, FITC labeled Fe 2 Tf; hTf, human serum transferrin; LSCM, laser scanning confocal microscopy; Tf, transferrin; TfR, transferrin receptor; ICP-AES, inductively coupled plasma atomic emission spectrometry. (Received 9 July 2002, revised 15 October 2002, accepted 21 October 2002) Eur. J. Biochem. 269, 6082–6090 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03326.x In the present work, we report a detailed study of uptake of Yb 3+ by human Tf (hTf) by UV-vis and NMR spectroscopy, and inductively coupled plasma atomic emission spectro- metry (ICP-AES). The binding of Yb-bound Tf to human erythroleukemia K562 cells was determined by flow cyto- metry and the cross-membrane transport was studied by confocal laser scanning microscopy (CLSM). EXPERIMENTAL PROCEDURES Materials Human Tf (catalog no. T0519), Hepes, Yb 2 O 3 ,Lu 2 O 3 , fluorescein isothiocyanate (FITC) and propidium iodide were purchased from Sigma. A Sephadex G-25 fine column was purchased from Pharmacia. K562 cells were obtained from the Laboratory of Immunology, Peking University. RPMI 1640, penicillin, streptomycin and fetal bovine serum were all purchased from Gibco. All the chemicals used were of analytical grade. All containers were rinsed with hydro- chloric acid to diminish the influence of metal ions. Ultrapure water was used to prepare solutions. Ytterbium chloride stock solution was prepared by dissolving ytterbium oxide in a minimum of concentrated HCl and adjusting pH to % 4 with NaOH. The solution was then diluted to the concentration needed prior to use. Ytterbium citrate was prepared by addition of 1 mol equiv of citrate to YbCl 3 solution and followed by pH adjustment to 4–6. Human erythroleukemia K562 cells were cultured in RPMI 1640 medium supplemented with 10% inactivated fetal bovine serum and 100 UÆmL )1 penicilin and strepto- mycin at 310 K in a 5% CO 2 atmosphere. All experiments wereperformedwithacelldensityof% 5 · 10 5 cells per mL. Preparation of Fe C -Tf and ferric-saturated Tf (Fe 2 -Tf) Monoferric- (selective loading of Fe 3+ in the C-lobe of Tf) and diferric-Tf were prepared as described previously [23]. The apo-Tf was purified by sequential dialysis against 0.1 M Hepes (pH 7.4) with 0.1 M NaClO 4 for 24 h and then against 0.1 M Hepes buffer pH 7.4 containing 0.1 M NaCl [24]. The addition of 1 or 2 mol equiv of Fe(nitrilotriace- tate) 2 to apo-TF and incubated the solution at 310 K for 30 min. The solution was then passed through Sephadex G- 25 column (1 · 10 cm) to remove the low molecular mass ligands (e.g. nitrilotriacetate), and Fe C -Tf or Fe 2 -Tf fraction was collected. Protein concentrations were determined spectrophotometrically on the basis of e 280 93 000, 103 000 and 113 000 M )1 Æcm )1 for apo-Tf, Fe C -Tf and Fe 2 -Tf, respectively [23]. The iron saturation of Fe C -Tf and Fe 2 -Tf was also estimated from the ratio of A 280nm /A 465nm [23]. The Yb-Tf solution was prepared by mixing YbCl 3 and apo-Tf solutions in a molar ratio of 2.5 : 1 and the free Yb was removed by ultrafiltration. As indicated in the Results section, with this molar ratio apo-Tf was not fully saturated by Yb 3+ however, the major Yb-bound Tf species is mainly Yb 2 -Tf. Fluorescence labeled Yb-Tf and Fe 2 -Tf were prepared by incubating Fe 2 -Tf and Yb 2 -Tf solution with FITC in 0.5 M Na 2 CO 3 -NaHCO 3 buffer (pH 9.0) at 277 K for 4–5 h [25]. The excess FITC was removed by gel filtration through a Sephadex G-25 column (1.0 · 10 cm). The green solution of FITC-labeled protein fraction was collected. The ratio of apo-Tf vs. FITC and the concentration of FITC-Tf were determined according to the literature [26]. Electronic absorption spectroscopy The apo-Tf solutions were prepared by diluting aliquots of a stock solution to % 1 · 10 )5 M )1 with 100 m M Hepes buffer (pH 7.4). Immediately before Yb 3+ was added, an aliquot of a concentrated solution of NaHCO 3 wasaddedtogivea 5m M bicarbonate solution. For titration experiments, aliquots of the stock solution of Yb 3+ wereaddedtothe apo-Tf cuvette and the spectra were recorded at intervals of at least 30 min at room temperature. Stopped-flow spectrophotometry The kinetic process of Yb 3+ binding to apo-Tf and Fe C -Tf was monitored on a SF18MV stopped-flow spectropho- tometer (Applied Photophysics). The changes in the absorption at 242 nm were followed upon mixing equal volume of the YbCl 3 and protein solutions (8.3 and 7.9 l M for apo-Tf and Fe C -Tf, respectively). The driving syringes were immersed in a water bath at 298 ± 0.5 K. Four- hundred data points were collected over various times (2, 5 and 10 s) for each trace and each curve was obtained by averaging 5–10 traces. The dead-time of the instrument under the experimental condition was less than 5 ms. Inductively coupled plasma atomic emission spectrometry These experiments were carried out on a Perkin-Elmer Plasmaquant 110 Emission Spectrometer using standard methods [27]. Yb-loaded proteins were prepared by addition of appropriate molar equivalent of metal ions to apo-Tf in 0.1 M Hepes buffer containing 0.1 M NaCl,atpH7.4,and left to incubate at room temperature for about 1 h. Then the samples were purified by using Centricon 30 (Amicon) ultrafilters, washed four times with 0.1 M Hepes buffer, followed by ultrafiltration after each washing. The final protein solutions were diluted with ultrapure water con- taining 1% nitric acid. The Yb content measured directly without digestion of the samples using ICP-AES at 369.3 nm. NMR spectroscopy 1 H NMR spectra were recorded at 500 MHz on a Bruker DRX500 spectrometer. Spectra were acquired using 0.6 mL of solution in a 5-mm tube at 298 K, % 1000 transients, 6-ls pulses, recycle time 1.8 s and 16 K data points. A solution of 0.8 m M apo-Tf in 0.1 M KCl, 10 m M NaHCO 3 was used for the 1 H NMR studies. The pH values of NMR solutions were kept at 7.40 ± 0.05 and was checked before and after NMR measurements with a Corning 440 pH-meter, equipped with an Aldrich microcombination electrode and calibrated with standard buffers at pH 4.0 and 7.0, respectively. Flow cytometry analysis The competitive effect of Yb-Tf species on binding of Fe 2 -Tf to K562 cells was studied by flow cytometry. The solutions with a constant concentration of FITC-labeled Fe 2 -Tf and Ó FEBS 2002 Ytterbium transferrin (Eur. J. Biochem. 269) 6083 various concentrations of unlabeled Yb-Tf (2.5 : 1) were each incubated with 1 mL of K562 cell suspension (10 6 cellsÆmL )1 )in10m M Hepes buffer (pH 7.4), contain- ing 0.15 M NaCl at 310 K for 30 min. After chilling (0 °C, 30 min) to terminate the reaction, the incubated mixtures were centrifuged at 500 g for 5 min and then the pellets were washed twice with ice-cold Hank’s balanced salt solution to remove the extracellular FITC-labeled Fe 2 -Tf. The fluorescence intensity of each single cell was recorded on a FASCAN flow cytometer (Becton-Dickinsin), with argon laser set at k ex 488 nm and k em 530 ± 26 nm. The results were expressed as the mean channel fluorescence intensity (  FF), which was obtained by integrating the intensity per cell from 1 · 10 4 cells for each sample. Laser scanning confocal microscopy Entry of FITC-Yb-Tf into K562 cells was investigated by LCSM. A suspension (1 mL) of 10 6 K562 cells was incubated with 1 l M FITC-Yb-Tf solution for 30 min at 310 K. The unbound FITC-Yb-Tf was removed by centrifugation with ice-cold 10 m M Hepes buffer at pH 7.4, containing 0.15 M NaCl. The cells were visualized under LCSM (TCS NT, Leica). Fluorescence images were collected at 512 · 512 pixels resolution using confocal microscopy with an oil immersion objective (apo-planar 40 · 1.25) in a Leica inverted microscopy, with k ex 488 nm for FITC-Yb-Tf or Fe 2 -Tf, and k ex 568 nm for propidium iodide. The emission fluorescence at k em 530/ 30 nm was collected with Band Pass (BP) filter and Long Pass (LP) 590 nm filter. RESULTS Binding of Yb 3+ to apo-hTf The thermodynamics of binding of Yb 3+ to apoTf. Binding of Yb 3+ to apo-Tf was investigated by addition of aliquots of Yb 3+ to a solution of apo-TF in 100 m M Hepes buffer in the presence of 5 m M NaHCO 3 . As shown in Fig. 1, upon addition of Yb 3+ to apo-Tf, two absorbance bands appeared at 242 and 292 nm, the characteristic of metal binding to phenolate groups of tyrosine residues at the specific iron bind-sites of apo-Tf. The De 242 increases linearly with the increase of Yb/apoTf ratio (r)uptoa value of r ¼ 1. The results indicated that Yb 3+ bound to tyrosine residues of the Fe 3+ -binding sites and induced deprotonation. From the initial slope of the curve, the molecular absorption coefficient (e)ofTfwithoneYb 3+ saturated was obtained as 26 500 ± 500 M )1 Æcm )1 . Above r ¼ 1, the absorption increases further but less profoundly, indicative of the occupation of a second site with a lower affinity. Under similar conditions, reactions of apo-Tf with Yb 3+ -citrate gave rise to the same two bands at 242 and 292 nm, respectively, and the same molecular absorption coefficient. We also noticed that the higher the bicarbonate concentration, the higher degree of saturation of trnasferrin by Yb 3+ (data not shown). The two dissociation constants (K d1 and K d2 )were determined based on the data in Fig. 1. According to Bjerrum equation [28], we have,  nn ¼ K 1 ½Yb 3þ  free þ 2K 1 K 2 ½Yb 3þ  2 free 1 þ K 1 ½Yb 3þ  free þ K 1 K 2 ½Yb 3þ  2 free ð1Þ where  nn ¼ ½Yb 3þ  bind ½Tf total , K 1 (¼ 1/K d1 )andK 2 (¼ 1/K d2 )are the stepwise association constants. At low [Yb 3+ ] total , all of the added Yb 3+ bind to apo ) Tf, [Yb 3+ ] bind ¼ [Yb 3 ±Tf ] and therefore,  nn ¼ ½Yb 3þ À Tf ½Tf total ¼ / ¼ A À A min A max À A min ð2Þ ½Yb 3þ  free ¼½Yb 3þ  total À½Yb 3þ  bind ¼½Yb 3þ  total À½Yb 3 Æ Tf ¼½Yb 3þ  total À /½Tf total ð3Þ By fitting the /$ [Yb 3+ ] free curve to Eqn (1) with those data shown in the insert of Fig. 1, we get the two dissociation constants K d1 ¼ 4.17 ± 0.4 l M and K d2 ¼ 18.5 ± 7.0 l M . ICP-AES was also used to study the binding ratio of Yb 3+ to apo-Tf. After addition of 2.0 and 2.5 mol equiv of Yb 3+ to apo-Tf, the final ratios of Yb 3+ to Tf, after removal of low molecular mass components via ultrafiltration, were % 1.70 : 1 and 1.98 : 1, respectively. This suggested that both the N- and C-lobe of apo-Tf could be occupied by Yb 3+ . The kinetics of binding of Yb 3+ to apo-Tf. The kinetics of Yb 3+ binding to apo-Tf was studied with the stopped-flow technique. The absorbance at 242 nm increased upon mixing Yb 3+ with apo-Tf and the rise was more rapid with increasing the molar ratio of Yb 3+ to apo-Tf (data not shown). The apparent rate constants were obtained from fitting the kinetic curves. The quality of the fits was assessed according to the residuals and the normalized variance (Fig. 2). The data can be fitted to the bi-exponential function (Eqn 4): fðxÞ¼A 1 expðÀR 1 xÞþA 2 expðÀR 2 xÞþc ð4Þ Where R 1 and R 2 are the rate constants of the two kinetic phases, and A 1 and A 2 are the corresponding amplitudes Fig. 1. UV difference spectra of apo-Tf (11.3 l M apo-Tf in 5 m M NaHCO 3 and 10 m M Hepes buffer at pH 7.4 and 310 K) after addition of various amounts of molar ratios of Yb 3+ (as YbCl 3 ). The two absorption bands at 242 and 295 nm are indicative of Yb 3+ binding to the specific iron sites of Tf. Molar ratio from bottom to top: 0.25, 0.50, 0.75, 1.0, 1.25, 1.5, 2.0, 3.0, 3.5 and 4.0. Inset: titration curve (242 nm) for Yb 3+ binding to apo-Tf. 6084 X l. Du et al. (Eur. J. Biochem. 269) Ó FEBS 2002 that show the contribution of individual kinetic phases to the observed change in the absorption. The molar absorp- tivity (e) and the normalized rate constant k can be obtained from the fitted values of A and R divided by the protein concentration (e ¼ A/[Tf] and k ¼ R/[Tf]). The depend- ence of e and k on the molar ratio of Yb 3+ to apo-Tf is shown in Fig. 3, upper and lower, respectively. In order to understand the two kinetic phases, Tf with iron selectively loaded into the C-lobe (Fe C -Tf) was prepared and the binding kinetics of Yb 3+ to Fe C -Tf was studied under the same experimental conditions. Quite differently, the kinetic data for the reaction of Yb 3+ and Fe C -Tf can readily be fitted to a single exponential function f(x) ¼ Aexp(–Rx) + c. The molar absorptivity (e)andthe normalized rate constant k can be obtained similarly and were shown in Fig. 2 for comparison. It was noticed that e 1 and e (13 300 ± 500 M )1 Æcm )1 ) were approximately the same and were only half the value of e 2 . The slope of the curve of rate constant k 1 vs. [Yb 3+ ]/protein (1.268 · 10 5 s )1 ) was again almost the same as that of k vs. [Yb 3+ ]/ protein (slope: 1.216 · 10 5 s )1 ). Citrate competition The competition between Tf and citrate for Yb 3+ binding was investigated by addition of small aliquots of citrate to the solution containing apo-Tf and 2.5 mol equivalents of Yb 3+ in the presence of 5 m M bicarbonate. The protein bound Yb 3+ (converted from DA 242 ) decreased almost linearly upon addition of 2, 4, 6, 8 and 10 mol equiv of citrate, and finally reached to < 10% of its original value. The very minor decrease in absorbance upon further addition of citrate was probably due to other factors (Fig. 4). Competition with ferric iron. When Fe 3+ wasaddedtoa solution of apo-Tf containing 2.5 mol equiv of Yb 3+ ,anew broad band centred at 465 nm appeared and increased in intensity gradually up to 2.0 mol equiv of Fe 3+ (data not shown). This indicated that 2 mol equiv of iron were sufficient to completely displace Yb 3+ from Yb-bound Tf species in solution of 2.5 : 1 mol ratio, suggesting that Fe 3+ binds to Tf more tightly than Yb 3+ (data not shown). The binding mode and the conformational changes induced by Lu 3+ and Yb 3+ : 1 H NMR studies These experiments were carried out in order to investigate the order of lobe loading and the conformational change induced by Yb 3+ . Because the signals of Yb 3+ -Tf systems were broadened, due to paramagnetic effect of Yb 3+ ,a parallel study on Lu 3+ and apo-Tf was conducted for comparison that will provide insight into the nature of Yb 3+ binding and the conformational changes in detail. A 1 H NMR spectrum of apo-Tf in the presence of 10 m M bicarbonate was recorded after addition of Yb 3+ (or Lu 3+ ) Fig. 2. Kinetics of binding of Yb 3+ to apo-Tf. The final concentration of the protein is 4.15 l M , and the molar ratio of Yb 3+ to apo-Tf is 1.75. Also shown are the residuals to the bi- (middle) and mono-exponential fits (the lower), suggesting that the bi-exponential function is a better fit. Fig. 3. The dependence of e (top) and k (bottom) on the molar ratio of Yb 3+ to apo-Tf. In the lower panel the data of k 1 , k 2 and k over the range of the ratio [Yb 3+ ]/protein between 2 and 6 are linearly fitted: k 1 ¼ 1.268x + 2.449, R 2 ¼ 0.979; k 2 ¼ 0.704x + 0.110, R 2 ¼ 0.952; k ¼ 1.216x + 0.508, R 2 ¼ 0.996. Ó FEBS 2002 Ytterbium transferrin (Eur. J. Biochem. 269) 6085 in steps of 0.5 mol equiv. Typical 1 H NMR spectra of apo- Tf in aliphatic and aromatic regions are shown in Figs 5 and 6. The resonance appeared in the high-field region (+0.5 to )1.0 p.p.m., Fig. 5A) was attributed to the methyl groups close to the surface of the aromatic rings of Phe, Trp and Tyr residues due to the ring current effects [29]. Signal b ()0.33 p.p.m) was related to the resonance of the CH 3 group of Leu122, which lies directly above the face of Trp128 [29–31]. Signals g and h were assigned to Leu122 and Val246, respectively; other signals were assigned to methyl groups in the C-lobe [29]. Following the addition of 1 mol equiv of Lu 3+ , signal c disappeared, signal d increased in intensity, while signal j decreased, and signal i shifted to the lower frequency. The other signals remained unchanged. This suggests that Lu 3+ binds to the C-lobe first. No changes appeared after further addition of Lu 3+ , which is similar to that of Bi 3+ [32]. More changes were observed in the 1 H NMR spectrum upon addition of Yb 3+ to Tf solution (Fig. 6). The typical 1 H NMR spectrum of Tf in the region of 2.0– 2.2 p.p.m. featured three sharp intense signals (Figs 5B and 6B), which are assigned to the N-acetyl moieties of AcNeu and GlcNac of the di-antennary glycan chains in the C-lobe [29]. Upon the addition of 1 mol equiv of Lu 3+ ,anew signal appeared at 2.097 p.p.m. (signal C), while the signal at 2.088 p.p.m. split into two (A and B). The relative intensities of the peaks altered upon further addition of Lu 3+ , but no shift occurred. Although the signals were broadened and shifted by Yb 3+ due to its paramagnetic character, the appearance of a new signal at 2.120 p.p.m. and the splitting of signals at 2.084 and 2.045 p.p.m. were still visible (Fig. 5B); Yb 3+ binding to the C-lobe can thus be inferred. These results suggested that Yb 3+ bind to the protein and probably altered the conformation of the protein in a manner similar to Fe 3+ . The signals in the region of 6.2–8.5 p.p.m. were usually attributed to His dCH resonance. As shown in Fig. 5C, when 0.5 mol equiv of Lu 3+ was added, signal p (6.360 p.p.m) decreased in intensity, while the new signal q appeared at 6.420 p.p.m. Further addition of Lu 3+ led to the disappearance of p and further increase of q, whereas a new signal r emerged at 7.75 p.p.m. No further changes were observed up to 1.5 mol equivalent Lu 3+ was added. In 2D-TOCSY 1 H NMR spectra, the signals at 6.34 p.p.m. (p) and 7.72 p.p.m. (r) were previously demonstrated to be correlated [32]. These changes induced by Lu 3+ indicated the disturbance on the microenvironment around His residues. In contrast, the signals of apo-Tf were severely broadened when Yb 3+ was added (data not shown). Binding of Yb 2 -Tf on K562 cell membrane Binding of FITC-Yb 2 -Tf to K562 cells. The binding of Yb-bound Tf species in solution with 2.5 : 1 mol ratio to K562 cellular membrane was quantitatively evaluated by sorting the cells after incubation with FITC-Yb 2 -Tf by flow cytometric technique based on the fluorescence excited at 488 nm. As shown in Fig. 7A, the mean channel fluorescence intensity (  FF), which reflects the amount of FITC-Yb-Tf species bound to the cells, was found to increase significantly at low concentration of FITC-Yb-Tf (<0.1l M Tf), probably due to specific binding of the protein to Tf receptor (TfR). When the concentration of FITC-Yb-Tf was over 0.1 l M , the increase in intensity became less profound and reached saturation at 0.2 l M . In contrast, the mean channel fluorescence intensity of Fig. 5. 1 H NMR Spectra of apo-Tf and its Lu 3+ complexes. (A) In the high-fieldregion()1.0 to +0.5 p.p.m), (B) N-acetyl region, and (C) in the aromatic region. Fig. 6. 1 H NMR spectra of apo-Tf and its Yb 3+ complexes in the high- field region ()1.0 to +0.5 p.p.m) (A) and N-acetyl region (B). Fig. 4. Competition between apo-Tf and citrate. Changes in percent of Yb bound to Tf (converted from the molar absorption coefficient at 242 nm) with increasing citrate concentration at 298 K and pH 7.4. 6086 X l. Du et al. (Eur. J. Biochem. 269) Ó FEBS 2002 FITC-Fe 2 -Tf increased more evidently and reached satura- tion at a much lower concentration (0.05 l M )thanthatof FITC-Yb-Tf species under the same conditions. This result confirmed the binding of Yb-Tf to K562 cellular membrane. The competitive effect of Yb 2 -Tf on the binding of Fe 2 -Tf to K562 cells. K562 cells were incubated with a constant concentration of FITC-Fe 2 -Tf solution in the presence of different molar ratios of Yb-Tf (2.5 : 1) at 273 K for 30 min. The cells were then sorted according to FITC fluorescence excited at 488 nm (Fig. 7). The results showed that upon the addition of Yb 2 -Tf and increasing the molar ratio of Yb-Tf /FITC-Fe 2 -Tf, the cell populations shift to lowered fluorescence (Fig. 7B), while the mean fluorescence intensity (  FF) decreased with increasing concentration of Yb- Tf (Fig. 7C). The  FF decreased to half of its original value when 18 mol equiv of Yb-Tf was incubated with cells in the presence of FITC-Fe 2 -Tf. Therefore part of the FITC-Fe 2 - Tf was inhibited from binding to TfR by Yb-Tf. These results suggested that Yb-bound Tf species compete with Fe 2 -Tf for the specific Fe 2 -Tf binding sites to the receptor on the cell membrane. The transport of FITC-Yb 3+ -apo-Tf into K562 cells observed by LCSM LSCM was employed to visualize the entry of FITC labeled Yb-bound Tf species into K562 cells. In Fig. 8A,B, the green fluorescence resulted from the emission of FITC labeled Yb-Tf when cells were excited at 488 nm. As the protein (Tf) was randomly labeled with FITC, its position and intensity represent the location and relative concentra- tion of the labeled Yb-Tf. The red fluorescence resulted from the emission of propidium iodide when the same cells were labeled with propidium iodide and excited at 560 nm. The appearance of green fluorescence inside K562 cells (Fig. 8C) clearly demonstrates that FITC labeled protein not only binds to K562 cellular membrane but also enters the cells. DISCUSSION There is an increasing interest in the use of lanthanide in medicine and biology [4–7]. 169 Yb has been considered as a potential replacement for 125 Iand 103 Pd. It was also chosen as a model nuclide for chemically related trivalent metal ions such as 90 Y 3+ and the clinically used 153 Sm 3+ [4,5]. However, not enough about transport proteins and the chemical structure in which 169 Yb is incorporated into cells is understood; in order to improve their accumulation in tumors, such knowledge is essential. Tf is a glycoprotein (80 kDa) present in blood plasma with a concentration of 35 l M .Itisonly30%saturatedwithironinbloodplasma and has the capacity for binding to other metal ions of therapeutic and diagnostic interest [16]. Therefore, it has been suggested that Tf can act as a nature ÔcarrierÕ for metallodrugs (e.g. 67 Ga, Ru and Ti anticancer agents) [33]. Complexation of metal ions to the phenolic group of the tyrosine residues in the specific iron site perturbs the p–p* transitions of the aromatic group and leads to two absorption bands at % 240 and 295 nm in the UV difference spectrum. The molar absorption coefficient (26 500 M )1 Æcm )1 ) upon binding of Yb 3+ to apo-Tf, is similar to other Ln to apo-Tf, for example, 20,400, 21 000 and 22 000 M )1 Æcm )1 for Lu 3+ ,Sm 3+ and Eu 3+ ,respect- ively [34]. Both UV and ICP-AES data suggested that two Yb 3+ bind to apo-Tf in the specific iron binding site and two tyrosines are involved in binding of Yb 3+ in both the N- and C-lobe as the case for Fe 3+ . The X-ray crystal structure of Sm 2 -lactoferrin revealed that Sm 3+ indeed binds to two tyrosine residues in both lobes of the protein and induces the Fig. 7. Saturation curve of binding of FITC-Fe 2 -Tf (filled square) and FITC-Yb-Tf (filled red circle) to the K562 cellular membrane (A), his- tograms of cellular fluorescence excited by 488 nm (B), and the inhibition of Yb-Tf on FITC-Fe 2 -Tf bound to K562 cells (C). The three curves in (B) represent the negative populations (curve ÔaÕ), cell populations after incubation with 0.48 M FITC-Fe 2 -Tf solution (curve ÔbÕ)andcell populations after incubation with solution with Yb 2 Tf: Fe 2 -Tf ¼ 18, 0.48 M FITC-Fe 2 -Tf (curve ÔcÕ). Ó FEBS 2002 Ytterbium transferrin (Eur. J. Biochem. 269) 6087 same overall structural changes in this protein as Fe 3+ [35]. However, binding of Yb 3+ to the strong site (the C-lobe) was in a molar ratio Yb 3+ /apo-Tf < 1: 1, while the binding to the other site is much weaker. To examine the kinetics of the uptake of Yb 3+ in the individual lobes, we carried out stopped-flow experiments and found two kinetic phases with twofold differences in rate constants. The almost identical molar absorptivity of e 1 and e and approximately the same slope of the k 1 and k (Fig. 3) suggests that thenatureofthefirstkineticphaseofthereaction between Yb 3+ and apo-Tf is the same as that of the reaction between Yb 3+ and Fe C -Tf. Therefore the rapid kinetic phase should correspond to the binding of the metal to the N-lobe, although the binding favorsthe C-lobe thermodynamically as judged by 1 H NMR. The more open conformation of the N- lobe may facilitate Yb 3+ ion binding [21]. Although 1 H NMR signals were considerably broadened upon addition of Yb 3+ to the protein due to paramagnetic properties, information can still be obtained by comparison with its analogue Lu 3+ . The sharp resonances at 2.0– 2.2 p.p.m. are attributable to the N-acetyl moieties of the glycan chains (NAcGlc and NacNeu residues) in the C-lobe of Tf. As these resonances were perturbed only on addition of the first mol equiv of Yb 3+ and Lu 3+ (Figs 5 and 6), this suggests preferential binding of Yb 3+ (and Lu 3+ )tothe C-lobe of Tf occurs. Similar behavior has been observed for several other metal ions [29,30,32]. The changes in shifts induced by progressive addition of Lu 3+ (and Yb 3+ )maybe due to the conformational change induced by metal ions, a common feature that was observed previously [16,29,30,32]. Our flow cytometric data (Fig. 7) clearly demonstrate that the Yb-bound Tf species can be recognized by and bind to TfR on the surface of K562 cells in a manner similar to Fe 2 -Tf. However, the higher saturation concentration of Yb 2 -Tf (0.2 l M vs. 0.05 l M for Fe 2 -Tf) and its competition with iron indicated a lower affinity of cell receptors for Yb 2 - Tf than for Fe 2 -Tf. LCSM offers an effective way to investigate the transport of extracellular molecules across the membranes and to identify the locations of the molecules within individual cells using appropriate fluorescent probes [36–38]. The green fluorescence inside cells indicated that FITC-Yb 2 -Tf entered cells and may have been located in cytoplasm when compared with cells labeled with propidium iodide. These results are in good agreement with flow cytometric study. However, they are merely the static pictures, reflecting the position of the Yb-bound Tf species at that moment. Further investigation is therefore needed to confirm whether Yb 3+ can be released from Yb-Tf species. Our data on the stability of the Yb 3+ –Tf complexes as a function of pH have shown that Yb 3+ was released under acidic conditions (pH < 6). Intracellular Yb has also been detected after incubation of Yb-Tf solution (mol ratio 2.5 : 1) with U87- MGcellsbyICP-MS(K.Wang,X.Du,Y.Chang,R.Li,J. Situ, H. Sun & Z. Qian, unpublished data). Thus, beside anion channel and other mechanism (e.g. citrate), the cellular uptake of Yb 3+ via Tf receptor-mediated endocy- tosis is highly possible. Different transport or uptake mechanisms seem to exist for the complexes of Yb 3+ and uptake of 169 Yb-citrate was previously shown to be an active cellular transport process. It is only dependent on the metabolic activity of the cells, however, and is not tumor specific [5]. The citrate concentra- tion in blood plasma (% 100 l M ) is comparable to Tf (35 l M ). Our citrate competition data indicated that binding of Yb 3+ to Tf is slightly stronger than to citrate (Fig. 4). The metal can be essentially removed from the protein in the presence of 10 mol equiv of citrate. Therefore, Tf is probably one of the (major) targets in blood plasma in addition to citrate. Previous in vivo andin vitro studies usinglabeled blood serum have suggested that the major target of Yb 3+ is Tf [12,13]. It is known that malignant cells have a higher iron requirement and subsequently express much higher Tf receptors [39]. Then the uptake of Yb-bound Tf by the tumor cells might be much higher than the normal cells and the uptake of iron by the tumor cells will be retarded much more than by the normal cells. Yb-Tf uptake mechanism via the receptor-mediated endocytosis is possible and this mechan- ism could selectively facilitate the accumulation of Yb 3+ in tumor cells. It is known that 67 Ga citrate, a commonly used radiopharmaceutical for soft tumors and abscess diagnosis, enters tumor cells via the Tf mediated endocytosis [11,40,41]. CONCLUSIONS The spectroscopic studies have shown that Yb 3+ binds to the two specific iron-sites of apo-Tf. The 1 HNMRdata show that Yb 3+ preferentially occupies the C-lobe rather than the N-lobe, as has been observed for several other metal ions and that Yb 3+ can be replaced by Fe 3+ . Interestingly, the binding to the N-lobe is kinetically favored. The binding of Yb 2 -Tf to K562 cell membranes (TfR) was demonstrated and was weaker than that of Fe 2 - Tf and only part of Fe 2 -Tf could be displaced. The confocal microscopic studies indicated that Yb 2 -Tf is likely to enter the cells in a way similar to Fe 2 -Tf, and at the same time almost without affecting the function of Tf. It will be interesting for the future to investigate whether Tf can enhance Yb 3+ accumulation in tumor cells. ACKNOWLEDGEMENTS This work was funded by National Natural Science Foundation of China (No. 29890280), the University of Hong Kong and the Hong Kong Polytechnic University. We thank the Area of Excellence Scheme of University Grants Committee (Hong Kong) for their support. Fig. 8. Transport of FITC-Yb-Tf into K562 cells visualized by LSCM. (A) K562 cells after incubation with FITC-Yb-Tf (B) propidium iodide-stained K562 cells and (C) merging the images. The green fluorescence indicated FITC- Yb 2 Tf entered into cells. 6088 X l. Du et al. (Eur. J. Biochem. 269) Ó FEBS 2002 REFERENCES 1. Smith, T., Shawe, D.J., Crawley, J.C.W. & Gumpel, J.M. (1988) Use of single photon emission computed tomography (SPECT) to study the distribution of 90 YinpatientswithBaker’s cysts and persistent synovitis of the knee. Ann. Rheum. Dis. 47, 553–558. 2. Canada, R.G. (1993) Calcium receptor binding of cisplatin and terbium in human breast tumor cells after hyperthermia. Radiat. Res. 133, 170–175. 3. Wang, K., Li, R.C., Cheng, Y. & Zhu, B. (1999) Lanthanides- the future drugs? Coord. Chem. Rev. 190–192, 297–308. 4. Lazarescu, G.R. & Battista, J.J. (1997) Analysis of the radio- biology of ytterbium-169 and iodine-125 permanent brachyther- apy implants. Phys.Med.Biol.42, 1727–1736. 5. Kampf, G., Knop, G., Franke, W.G., Bergmann, R. & Johannsen, B. (1997) Uptake of 169 Yb complexes in normal and tumour cells: influence of ligand and metabolic cell activity and stability of cellular association. Nucl. Med. Biol. 24, 349–355. 6. Evans, C.H. (1990) Biochemistry of the Lanthanides.PlenumPress, New York. 7. Cheng, Y., Huo, Q.H., Lu, J.F., Li, R.C. & Wang, K. (1999) The transport kinetics of lanthanide species in a single erythrocyte probed by confocal laser scanning microscopy. J. Biol. Inorg. Chem. 4, 447–456. 8. Shastri, B.P., Sankaram, M.B. & Easwaran, K.R. (1987) Car- boxylic ionophore (lasalocid A and A23187) mediated lanthanide ion transport across phospholipid vesicles. Biochemistry 26, 4925– 4930. 9. Powis, D.A., Clark, C.L. & O’Brien, K.J. (1994) Lanthanum can be transported by the sodium-calcium exchange pathway and directly triggers catecholamine release from bovine chromaffin cells. Cell Calcium 16, 377–390. 10. Cheng,Y.,Liu,M.Z.,Li,R.C.,Wang,C.,Bai,C.L.&Wang,K. (1999) Gadolinium induces domain and pore formation of human erythrocyte membrane: an atomic force microscopic study. Bio- chim. Biophys. Acta 1421, 249–260. 11. Jackson, G.E. & Byrne, M.J. (1996) Metal ion speciation in blood plasma: gallium-67-citrate and MRI contrast agents. J. Nucl. Med. 37, 379–386. 12. Taylor, D.M., Duffield, J.R., Williams, D.R., Yuele, L., Gaskin, P.W. & Unalkat, P. (1991) Binding of f-elements to the iron- transport protein transferrin. Eur. J. Solid. State. Inorg. Chem. 28, 271–274. 13. Meng, L., Ding, L., Chen, H.T., Zhao, D.Q. & Ni, J.Z. (1999) Biologicaleffectofrareearth(I)-contentanddistributionofrare earth in normal human plasma. Chem.J.Chin.University20, 5–8. 14. Luk, K.C. (1971) Study of the nature of the metal-binding sites and estimate of the distance between the metal-binding sites in transferrin using trivalent lanthanide ions as fluorescent probes. Biochemistry 10, 2838–2843. 15. Aisen, P. (1998) Transferrin, the transferrin receptor, and the uptake of iron by cells. Metal Ions Biol. Syst. 35, 585–631. 16. Sun, H., Li, H. & Sadler, P.J. (1999) Transferrin as a metal ion mediator. Chem. Rev. 99, 2817–2842. 17. Harris, A.W. & Sephton, R.G. (1977) Transferrin promotion of 67 Ga and 59 Fe uptake by cultured mouse myeloma cells. Cancer Res. 37, 3634–3638. 18. van Leeuwen-Stok, A.E., Schuurhuis, G.J., Drager, A.M., Visser- Platier, A.W., Teule, G.J. & Huijgens, P.C. (1996) Effect of modulation of the transferrin receptor on gallium-67 uptake and cytotoxicity in lymphoma cell lines. Br.J.Cancer74, 619–624. 19. Roskams, A.J. & Connor, J.R. (1990) Aluminum access to the brain: a role for transferrin and its receptor. Proc.NatlAcad.Sci. USA 87, 9024–9027. 20. Zuccola, H.J. (1993) The crystal structure of monoferric human serum transferrin. PhD Thesis, Georgia Institute of Technology, Atlanta, GA. 21. Jeffrey, P.D., Bewley, M.C., MacGillivray, R.T.A., Mason, A.B., Woodworth, R.C. & Baker, E.N. (1998) Ligand-induced con- formational change in transferrins: crystal structure of the open form of the N-terminal half-molecule of human transferrin. Bio- chemistry 37, 13978–13986. 22. MacGillivray, R.T.A., Moore, S.A., Chen, J., Anderson, B.F., Baker, H., Luo, Y., Bewley, M., Smith, C.A., Murphy, M.E.P., Wang, Y., Mason, A.B., Woodworth, R.C., Brayer, G.D. & Baker, E.N. (1998) Two high-resolution crystal structures of the recombinant N-lobe of human transferrin reveal a structural change implicated in iron release. Biochemistry 37, 7919–7928. 23. Martin, D.M., Chasteen, N.D. & Grady, J.K. (1991) Fluorescence and kinetic properties of Ru (III) (NH 3 ) 5 modified transferrin. Biochim. Biophys. Acta 1076, 252–258. 24. Bali, P.K. & Harris, W.R. (1990) Site-specific rate constants for iron removal from diferric transferrin by nitrilotris (methylene- phosphonic acid) and pyrophosphate. Arch. Biochim. Biophys. 281, 251–256. 25. Musgrove, E., Rugg, C., Tayor, I. & Hedley, D. (1984) Transferrin receptor expression during exponential and plateau phase growth of human tumour cells in culture. J. Cell Physiol. 118, 6–12. 26. Aladar, J. & Kalman, K. (1966) Chemical characterization of fluorescein isothiocyanate-protein conjugates. Biochim. Biophys. Acta 124, 166–175. 27. Zhang, L., Szeto, K.Y., Wong, W.B., Loh, T.T., Sadler, P.J. & Sun, H. (2001) Interactions of bismuth with human lactoferrin and recognition of the Bi(III)-lactoferrin complex by intestinal cells. Biochemistry 40, 13281–13287. 28. Naik, D.V., Paul, W.L., Threatte, R.M. & Schulman, S.G. (1975) Fluorometric determination of drug–protein association con- stants: the binding of 8-anilino-1-naphthalenesulfonate by bovine serum albumin. Anal. Chem. 47, 267–270. 29. Kubal, G., Mason, A.B., Sadler, P.J., Tucker, A. & Woodworth, R.C. (1992) Uptake of Al 3+ into the N-lobe of human serum transferrin. Biochem. J. 285, 711–714. 30. Kubal, G., Mason, A.B., Patel, S.U., Sadler, P.J. & Woodworth, R.C. (1993) Oxalate- and Ga(3+)-induced structural changes in human serum transferrin and its recombinant N-lobe. 1 HNMR detection of preferential C-lobe Ga 3+ binding. Biochemistry 32, 3387–3395. 31. Kubal, G., Sadler, P.J. & Tucker, A. (1994) pH-induced structural changes in human serum apotransferrin. pKa values of histidine residues and N-terminal amino group determined by 1 H-NMR spectroscopy. Eur. J. Biochem. 220, 781–787. 32. Li, H., Sadler, P.J. & Sun, H. (1996) Unexpectedly strong binding of a large metal ion (Bi3+) to human serum transferrin. J. Biol. Chem. 271, 9483–9489. 33. Li, H., Sun, H. & Qian, Z.M. (2002) The role of the transferrin- transferrin-receptor system in drug delivery and targeting. Trends Pharmacol. Sci. 23, 206–209. 34. Harris, W.R., Yang, B., Abdollahi, S. & Hamada, Y. (1999) Steric restrictions on the binding of large metal ions to serum transferrin. J. Inorg. Biochem. 76, 231–242. 35. Sharma, A.K. & Singh, T.P. (1999) Lactoferrin–metal inter- actions: first crystal structure of a complex of lactoferrin with a lanthanide ion (Sm 3+ )at3.4A ˚ resolution. Acta Crystallogr. D55, 1799–1804. 36. Kitamura, T., Gatmaitan, Z. & Arias, I.M. (1990) Serial quanti- tative image analysis and confocal microscopy of hepatic uptake, intracellular distribution and biliary secretion of a fluorescent bile acid analog in rat hepatocyte doublets. Hepatology 12, 1358–1364. Ó FEBS 2002 Ytterbium transferrin (Eur. J. Biochem. 269) 6089 37. Kotze ´ , A.F., Lueben, H.L., de Leeuw, B.J., de Boer, B.G., Ver- hoef, J.C. & Jungier, H.E. (1997) N-trimethyl chitosan chloride as a potential absorption enhancer across mucosal surfaces: In vitro evaluation in intestinal epithelial cells (Caco-2). Pharmaceut. Res. 14, 1197–1202. 38. Drevets, D. & Campbell, P.A. (1991) Macrophage phagocytosis: use of fluorescence microscopy to distinguish between extracellular and intracellular bacteria. J. Immunol. Meth. 142, 31–38. 39. Flulk, W.P., His, B.L. & Stevens, P.J. (1980) Transferrin and transferrin receptors in carcinoma of the breast. Lancet 2,390– 392. 40. Bernstein, L.R. (1998) Mechanisms of therapeutic activity for gallium. Pharmacol. Rev. 50, 665–682. 41. Harris, W.R. (1998) Binding and transport of nonferrous metals by serum transferrin. Struct. Bonding 92, 122–162. SUPPLEMENTARY MATERIAL The following material is available from http://www.black- wellpublishing.com/products/journals/suppmat/EJB/ EJB3326/EJB3326sm.htm Fig. S1. The effect of pH on the binding of Yb 3+ to apo-Tf. 6090 X l. Du et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Complexation of ytterbium to human transferrin and its uptake by K562 cells Xiu-lian Du 1 , Tian-lan Zhang 1 , Lan Yuan 1 , Yong-yuan Zhao 1 ,. detailed study of uptake of Yb 3+ by human Tf (hTf) by UV-vis and NMR spectroscopy, and inductively coupled plasma atomic emission spectro- metry (ICP-AES).