THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 258, No. 16, Issue of August 25, pp. 9681-9689, 1983 Printed in U.S.A.

Kinetics of Internalization and Recycling of Transferrin and the Transferrin Receptor in a Human Hepatoma Cell Line EFFECT OF LYSOSOMOTROPIC AGENTS*

(Received for publication, February 28, 1983)

Aaron CiechanoverS, Alan L. Schwartzgll, Alice Dautry-Varsat 11, and Harvey F. Lodish From the Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 and §Division of Pediatric HematologylOncology, Children’s Hospital Medical Center, Sidney Farber Cancer Institute and Department of Pediatrics, Haruard Medical School, Boston, Massachusetts 021 15

Growing HepG2 cells contain 50,000 functional sur- face transferrin-binding sites (Ciechanover, A., Schwartz, A. L., and Lodish, H. F. (1983) Cell 32,267- 275) and 100,000 intracellular sites. At saturating concentrations of [68Fe]transferrin, and under condi- tions in which protein synthesis is blocked, iron uptake is linear for several hours at a rate of 9,500 transferrin molecules/cell/min. Thus, each receptor must recycle a ligand, on the average, each 15.8 min. Surface-bound transferrin is rapidly endocytosed (t l i2 = 3.5 min). All of the iron remains within the cell, while the apotrans- ferrin is rapidly ( t I l2 = 5.0 min) secreted into the me- dium. Previously, we showed (Dautry-Varsat, A., Cie- chanover, A., and Lodish, H. F. (1983) Proc. Nutl. Acad. Sci. U. S. A. 80,2258-2262) that exposure of a ferrotransferrin-receptor complex to medium of pH less than 5.0 results in dissociation of iron, but that apotransferrin remains bound to its receptor. If the pH is raised to 7.0, such as would occur when an acidic intracellular vesicle fuses with the plasma membrane, apotransferrin is very rapidly dissociated (t1,2 = 17 s at 37 “C) from its receptor. Taken together, these re- sults indicate that transferrin remains bound to its receptor throughout the endocytic cycle. In the present study, we have directly measured all the kinetic pa- rameters involved in the transferrin receptor cycle. They are similar to those of the asialoglycoprotein receptor in the same cell line, and can be described by a simple kinetic model.

In the presence of lysosomotropic agents, ferrotrans- ferrin binds to its surface receptor and is internalized normally. However, iron is not dissociated from trans- ferrin, and ferrotransferrin recycles back to the cell surface and is secreted into the medium. We conclude that the low pH in endocytic vesicles is essential for the dissociation of iron from transferrin and its deliv- ery to the cell, but is not required for recycling of transferrin, and presumably of its receptor.

* This study was supported by Grant GM27989 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by the Melvin Brown Memorial Fellowship through the Israel Cancer Research Fund and by fellowships from the Leu- kemia Society of America and the Medical Foundation, Inc.

7 Supported in part by the John and George Hartford Foundation and the National Foundation, March of Dimes, Dad O’Connor Research Grant.

11 On leave from the Pasteur Institute, Paris, and the recipient of a fellowship from the Philippe Foundation.

During receptor-mediated endocytosis, a ligand binds to a specific cell surface receptor, and the receptor-ligand complex is internalized in clathrin-coated pits and vesicles. The fate of endocytosed transferrin is different from that of all other known ligands. Most ligands dissociate from the receptor in a prelysosomal vesicle ( l ) , most probably due to the low pH (2, 3). Ligands are then transported in a still unknown way to lysosomes, wherein they are degraded. In most of the systems studied, the receptor is not degraded, but recycles to the cell surface. Evidence for receptor recycling has been obtained for the asialoglycoprotein (4-6), mannose 6-phos- phate (7), mannose (8), low density lipoprotein (9), cy2-

macroglobulin (lo), and insulin receptors (1 1). Transferrin is a serum glycoprotein that transports iron

from sites of absorption and storage to tissue cells. After endocytosis, the transferrin polypeptide is not degraded, but rather is exocytosed intact into the medium; the two bound iron atoms remain within the cell (12-14). In this paper, we describe in detail the kinetics of internalization and exocytosis of transferrin and its receptor in the human hepatoma cell line HepG2. By two independent methods, we show that the cycle time of each functional receptor is about 16 min. Sur- face-bound ferrotransferrin molecules are internalized rapidly ( tI l2 = 3.5 min) and iron-free transferrin quickly ( tlI2 = 5.0 min) returns to the cell surface, apparently bound to its receptor, and is released into the medium.

Our recent studies (15) have partially clarified the role of the low pH of endocytic vesicles in recycling of transferrin. At pH values less than 5, iron is dissociated from the trans- ferrin-receptor complex, but apotransferrin remains bound to its receptor. This behavior contrasts with that of receptors for insulin, low density lipoproteins, and asialoglycoproteins, where ligand dissociates from the receptor at pH values less than 5 . When the pH is raised to 7, such as would occur upon fusion of an acidic intracellular vesicle with the surface mem- brane, apotransferrin is instantly dissociated from the recep- tor. A similar model based on kinetic studies in K562 cells was proposed by Klausner and co-workers (3, 14). By making use of weak bases which raise the pH of the endocytic vesicles (16, 17), we have studied the effect of the acidic pH on the dissociation of iron, and on the recycling of the receptor and transferrin to the cell surface.

EXPERIMENTAL PROCEDURES

Cells-All experiments were conducted on the human hepatoma line HepGP (18), cultured as described previously (19).

Materiak-Iron-free human transferrin and pronase were pur- chased from Calbiochem, and Eagle’s minimal essential medium and fetal bovine serum were from Gibco. NaXz5I and 59FeC13 were obtained

968 1

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

9682 Transferrin Receptor Recycling

from Amersham. All other chemicals were of the highest reagent grade available.

Labeling of Transferrin-Transferrin was saturated with iron and iodinated with '"1 as described elsewhere (13). For loading transferrin with 59Fe, 0.3 pmol of 59FeC12 was neutralized with NaOH and 6 pmol of nitrilotriacetic acid were added. Following neutralization, boiling for 5 min under nitrogen, and cooling on ice, the solution (0.45 ml) was added to 0.135 pmol of iron-free transferrin dissolved in 50 pl of 0.1 M NaClO,, 10 mM Tris-HC1, and 20 mM NaHC03, pH 7.6. After 1-h incubation at room temperature, the labeled protein was separated from free iron by chromatography on a Sephadex G-25 column (1 X 50 cm), followed by dialysis against 20 mM Na-HEPES,' pH 7.3. The specific activity was about 1.2 X lo6 cpm/nmol. Based upon the specific activity of the "Fe and protein concentration, the molar ratio of Fe/protein was 1.6 to 1.91 in different preparations.

Binding Assays-All binding assays using 1251-ferrotransferrin were carried out in tissue culture dishes (35 mm, Falcon) which were generally seeded 5 days prior to assay a t 0.5-1 X lo6 cells/dish. At the time of assay, the dishes were near confluence (2-4 X lo6 cells/ dish). [s9Fe]Transferrin-binding assays were carried out similarly, but in 60-mm dishes. The cells were washed twice a t 4 "C with Eagle's minimal essential medium buffered with 20 mM Na-HEPES, pH 7.3, and gassed with 95% 0, and 5% CO, (binding medium), followed by incubation for 30 min at 37 "C in this medium. Unless otherwise stated, binding was carried out in duplicate dishes a t 4 "C for 2 h in binding medium containing 77 nM labeled transferrin. Nonspecific binding was determined in the presence of a 100-fold excess of unlabeled ligand, and ranged from 2 to 15% of the specific binding. (The nonspecific binding increased with time after radiolabeling, probably due to inactivation or denaturation of the ligand.) All specific binding data were corrected accordingly. After washing off excess unlabeled ligand and dissolving the cells in NaOH as described earlier (19), the radioactivity in cell-associated ligand was measured in a Packard Auto-Gamma 500 y counter.

When metabolic inhibitors were used, binding was conducted in Hanks' balanced salt solution buffered with 25 mM Na-HEPES, pH 7.3. The extent of ligand binding in this solution was the same as in standard binding medium.

Internalization of Ligand As Measured by Treatment of Cells with Pronase-This procedure has been described elsewhere (13). Briefly, after binding and washing off unbound labeled ligand, the cells were incubated with 1 ml of binding medium containing 128 nM unlabeled transferrin, prewarmed to the appropriate temperature. At the indi- cated times, the medium was quickly removed, and the cells imme- diately were chilled by immersion in ice-cold phosphate-buffered saline (containing 1.7 mM CaC12), followed by 1-h incubation a t 4 "C in 1 ml of 0.25% pronase in binding medium. At the end of the incubation, cells were completely detached from the dish by repeated pipetting, and centrifuged for 1 min in an Eppendorf microfuge. The radioactivity in the supernatant and the pellet was determined.

Binding in Single-cell Suspension-A single-cell suspension was obtained by treatment of monolayer cells with EDTA. Following preincubation at 37 "C as described above, 1 ml of binding medium containing 10 mM EDTA was added to each 35-mm dish. The cells were gently scraped and pipetted free from the culture dishes and used directly for binding of labeled transferrin. Following binding, the cells were recovered by centrifugation at 800 X g for 5 min at 4 "C and were washed three times with 2 ml of binding medium containing 128 nM unlabeled ligand. The pellet was dissolved and counted as described above.

Binding of Transferrin to Receptors in Detergent-solubilized Cells- The cells in a 35-mm dish were washed as described above, and dissolved in 0.5 ml of binding medium containing 1% Nonidet P-40. Following incubation with '251-ferrotransferrin a t 4 "C for 2 h, 0.5 ml of a solution of 65% ammonium sulfate, 100 mM Tris-HC1, pH 7.6, was added to the mixture, and the samples were kept on ice for 10 min before filtration through a 24-mm Whatman GFC glass fiber disc. The filters were washed five times with 3 ml each of 32.5% ammonium sulfate, 50 mM Tris-HC1, pH 7.6, 1% Nonidet P-40, 1 mg/ml bovine serm albumin, and 0.1 mg/ml unlabeled transferrin, and were counted as described above.

Protein Determination-Protein was assayed according to Lowry et al. (20), using bovine serum albumin as standard.

' The abbreviations used are: HEPES, N-2-hydroxyethylpipera- zine-N'-Z-ethanesulfonic acid; Tf . Fe, ferrotransferrin; Tf, apotrans- ferrin.

Kinetic Model for Receptor Function and Recycling-The kinetic model in Fig. 1 depicts the essential parameters of the cycling of transferrin. It is similar to, but more extensive than, the one developed by Wiley and Cunningham (21) for epidermal growth factor receptor and by Schwartz et al. (5) for the asialoglycoprotein receptor. That, because the transferrin receptor can be directly followed throughout the successive steps of the endocytic cycle by following the labeled ligand, and most of the parameters can be determined by labeling either the apo- or the coligand. kl is the rate constant for binding of ferrotransferrin to unoccupied cell surface receptor (I?)*, and k-, is the rate constant for dissociation of ferrotransferrin from the cell surface receptor. k , is the first order rate constant for internalization of the receptor-ligand complex (R-Tf.Fe),. ks is the rate constant for dissociation of iron from the internalized receptor-ferrotransferrin complex (R-Tf .Fe),, and k, is the first order rate constant for move- ment of the intracellular receptor-apotransferrin complex, (R-Tf),, to the surface, thus generating a surface apotransferrin-receptor com- plex, (R-To.. b is the first order rate constant for dissociation of apotransferrin from the receptor and regeneration of unoccupied cell surface receptor, (R)a. In practice, we were unable to separate k3 from k,, and thus we employ a single rate constant, k,, to encompass both reactions. By definition kx-' = ks-l + k4-'. Thus, we employ the term (R-Tf), to denote all intracellular receptor-transferrin complexes, with or without iron.

In the formulation of this model, we have made several simplifying assumptions, all of which have been validated by experiments. We assume that all functional receptors behave identically, and that there is no alteration in functional receptor number during short periods of incubation. In particular, there is little synthesis or degradation of receptor molecules during this period. We assume that, in the steady state, there is no intracellular pool of receptors not bound to trans-

I \ \

Fe (Storage)

FIG. 1. A kinetic model for receptor-mediated endocytosis of transferrin. TI. Fe, ferrotransferrin. (R)s, unoccupied surface receptor. (R-TI.Fe),, surface ferrotransferrin-receptor complex. (R- TI.Fe)i, intracellular ferrotransferrin-receptor complex. (R-Tt),, in- tracellular apotransferrin-receptor complex. (R-T,)., surface apo- transferrin-receptor complex. TI, apotransferrin. kl, rate constant for binding of ferrotransferrin to surace receptors. k-', rate constant for dissociation of ferrotransferrin from cell surface receptors. k,, rate constant for internalization of surface ferrotransferrin-receptor com- plex. k3, rate constant for dissociation of iron from internalized ferrotransferrin-receptor complexes. k4, rate constant for movement of the receptor-apotransferrin complex to the surface. k,, overall rate constant for dissociation of iron and movement of the apotransferrin- receptor complex to the cell surface. b, rate constant for dissociatioz of apotransferrin from cell surface receptor.

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

Transferrin Receptor Recycling 9683

ferrin. Thus, the total number of receptors is constant, and is com- posed of unoccupied surface receptors (R)# , occupied surface receptor (R-Tf.Fe), and (R-Tf),, and internal receptors (R-Tf),. Therefore:

We can write:

" d ( R ) s - kl(R-Tf.Fe), - kl(Tf.Fe) ( R ) , + ko (R-Tf), (2) d t

d(R-Tf.Fe)s dt

= kl(Tf.Fe)(R), - (k-l + k ~ ) (R-Tf.Fe), (3)

d(R-Tf)L ~- - k,(R-Tf.Fe). - k,(R-Tf), d t

Since the concentration of Tf.Fe in the medium does not change significantly over the time of incubation, k,(Tf. Fe) can be written as one rate constant (klTf.Fe) or ( k l L ) . The half-times, tl/,, for each of the above reactions are equal to In 2/k. However, the mean time the receptor spends in each state is equal to tl/,/ln 2 or to k-'. Thus, the mean time required for the surface receptor ( R ) , to bind a ligand is equal to (k, .L)", and the mean time required for a receptor to traverse the entire endocytic cycle, T, is defined by:

T, = (k,L)" + (k,)" + ( k J 1 + (ko)-I (6)

Under "Results," we describe experiments which validate the above assumptions and allow us to directly determine the values of each of the above rate constants and cycle times, and also the proportion of the different functional forms of the transferrin receptor during steady state endocytosis.

RESULTS

The Fate of Transferrin Polypeptide and Iron during a Single Cycle of Endocytosis-Our first experiments focus on the fate of the protein and iron moieties of transferrin during a single cycle of endocytosis in HepG2 cells. In these studies, a saturating amount of 12'1- or [59Fe]transferrin is bound at 4 "C to the surface of HepG2 cells. Unbound ligand is re- moved, and the cells are incubated for various times at 37 "C. The medium is quickly removed, and the cells are chilled and treated with pronase for 1 h a t 4 "C. Only surface-bound ligand is accessible to the proteolytic enzyme, whereas inter- nalized ligand is protected from proteolysis and is recovered with the cell pellet. At least 39% of surface-bound lZ51-trans- ferrin is internalized within 5 min, and then is exocytosed into the medium (Fig. 2 ) . All of the exocytosed lZ51 radioactiv- ity is in intact transferrin, as shown by sodium dodecyl sulfate gel electrophoresis (data not shown). Similar results have been reported by others (12-14).

The fate of the iron moiety of transferrin is different. As can be seen from Fig. 3, 63% of the iron of surface-bound ["Feltransferrin is subject to endocytosis and remains within the cell. 37% of the 59Fe is released into the medium as intact [59Fe]transferrin (data not shown).

Dissociation of Transferrin from Surface Receptors at 37 "C-The simplest interpretation of these experiments is that about 37% of surface-bound transferrin is dissociated directly into the medium, while the balance is subjected to endocytosis. In particular, all of the iron subject to endocytosis remains in the cell, while all of the transferrin subject to endocytosis is exocytosed. To establish this point, it is nec- essary to measure the rate of dissociation of ferrotransferrin from surface receptors a t 37 "C. This is accomplished by using two energy inhibitors, NaCN and 2-deoxy-D-glucose. To- gether, these inhibitors completely block endocytosis of the labeled ligand into the cell, and all the cell-associated radio- activity is accessible to extracellular proteolytic eI.zymes (data

/Medium 0

/

0" 2.5 10 20 30 40 50 TIME ( M I N )

FIG. 2. Diacytosis of '2SI-transferrin in HepG2 cells. Trans- ferrin was bound to HepG2 cells at 4 "C as described under "Experi- mental Procedures." After washing off excess unbound ligand, the cells were incubated at 37 "C following addition of prewarmed binding medium containing 128 nM unlabeled transferrin. At the indicated times, the medium was quickly removed and the cells were chilled in ice-cold phosphate-buffered saline (containing 1.7 mM CaC1,) and treated with pronase as described under "Experimental Procedures." The radioactivity in the pronase-resistant fraction (O), pronase- sensitive fraction (0), and the medium (0) was determined.

1.0

0.1 - 0 6 10

TIME (MIN)

0 10 20 30 60 120 240 TIME (MIN)

FIG. 3. Single cycle receptor-mediated endocytosis of ["Fe] transferrin in HepG2 cells. The experimental details are as delin- eated in Fig. 2 except that [6gFe]transferrin was bound to the cells instead of 1251-transferrin. Inset, the accumulation of 59Fe in the cell and the dissociation of [59Fe]transferrin into the medium were re- plotted semilogarithmically as 1 - (BL/Bmax) versus time, where Bmax is the maximum amount in the compartment, and B, is the amount at time t. From the intracellular values, 8% was subtracted to correct for the background of pronase-resistant radioactivity at zero time.

not shown). ATP depletion is reversible, and when the inhib- itors are removed, the ligand is taken up by the cell and becomes resistant to extracellular pronase (not shown). To measure the rate of dissociation of ferrotransferrin, lZ5I-fer- rotransferrin is bound to cells a t 4 "C in the presence of

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

9684 Transferrin Receptor Recycling

inhibitors. After washing off excess unbound ligand, Hanks' solution, prewarmed to 37 "C and containing the metabolic inhibitors and 128 nM unlabeled transferrin, is added. The monolayers are then incubated at 37 "C (Fig. 4). In the pres- ence of energy inhibitors, '251-transferrin dissociates from surface receptors in a first order process with a half-time of 7.7 min. Thus, k-I = 0.090 min" (Fig. 4).

The Rate of Binding of '"I-Ferrotransferrin to Surface Re- ceptors at 37 "C-The rate of binding of '"I-ferrotransferrin to surface receptors at 37 "C is measured in the presence of the two inhibitors of ATP generation. At 6 pg/ml (77 nM) ferrotransferrin, the half-time for binding is about 3.0 min, compared to 21 min at 4 "C (Fig. 5 ) . In the presence of the energy inhibitors and a t 37 "C, the maximum amount of binding is similar to that obtained at 4 "C, confirming that there is no internalization of ligand. This level is similar to that obtained at 4 "C without inhibitors, indicating that the inhibitors do not affect the interaction between the ligand and the cell surface receptor (6). At 37 "C, the overall binding rate constant ( K I L ) is therefore In 2/3.0 = 0.23 min-'.

Rate of Internalization of Surface Transferrin-Receptor Complex-The rate of internalization is equivalent to the rate of acquisition of pronase resistance of a ligand prebound at 4 "C, upon warming to 37 "C. Since the iron moiety remains within the cell, we utilized ["Feltransferrin for this study. The half-time for internalization is 2.3 min (Fig. 3). kz is therefore 0.30 min".

This rate constant can also be calculated from the rate of loss of lZ"I-ferrotransferrin from the cell surface. This is the sum of two distinct processes, internalization and dissocia- tion:

- d(R-Tf.Fe)' = ( k-' + k,) (R-Tf.Fe), dt (7)

- W 3 -J d

e Control f\ c,

- v)

N TIME (MIN)

FIG. 4. Time course of loss of 1Z61-ferrotransferrin from the surface of HepG2 cells. '251-Ferrotransferrin was bound to cells as described under "Experimental Procedures" in the absence (0) or in the presence (0) of 10 mM NaCN and 20 mM 2-deoxy-~-glucose. After washing off excess unlabeled ligand, the cells were incubated at 37 "C after adding the appropriate prewarmed binding medium con- taining 128 nM unlabeled transferrin with or without energy inhibi- tors. At the indicated times, the medium was quickly removed. When energy inhibitors were present, radioactivity associated with the cells was directly measured, while in the absence of the inhibitors, the cells were quickly chilled in ice-cold phosphate-buffered saline (containing 1.7 mM CaC1,) and were treated with pronase as described under "Experimental Procedures." The pronase-sensitive fraction is plotted semilogarithmically uersus time. The data for the control experiment (i.e. that remaining on the cell surface) are adapted from Fig. 2, and were recalibrated after subtraction of a background of 6% that rep- resents surface-bound transferrin radioactivity which cannot be re- moved by pronase.

FIG. 5. Time course of ""I-ferrotransferrin binding to HepG2 cells in the presence of energy inhibitors. Iz5I-Ferro- transferrin (77 nM) was bound to HepG2 cells for the indicated times as described under "Experimental Procedures" at 4 "C and at 37 "C in the presence of 10 mM NaCN and 20 mM 2-deoxy-D-glucose. Inset, semilogarithmic replotting of the data as described in Fig. 3.

or t1I2 (loss from surface) = In 2/( k-, + kz) . To measure ( k-' + k 2 ) , we follow the fraction of '"1-ferrotransferrin, prebound to cells at 4 "C, that remains on the cell surface following warming of the cells to 37 "C (Fig. 2; Fig. 4, Control). From a semilogarithmic plot of the data, we find the half-time to be 2.4 min. k p is therefore equivalent to In 2/t1/2 - k-l = 0.29 - 0.090 = 0.20. The difference between the two values of kz, 0.20 and 0.30 min-', is due to experiment-to-experiment var- iability. In the remainder of this paper, we use a value of kz = 0.20 min-', the mean value obtained from several repeats of these experiments.

Knowing k-l and kZ, we can calculate the expected ratio between the internalization and direct dissociation of surface- bound transferrin molecules: k 2 / ( k z + k-,) = 0.69. From Fig. 3, we see that about 63% of the transferrin molecules initially bound to the cell surface are internalized, close to the expected value. This result also shows that both iron atoms are released in a single endocytic cycle.

Rate of Appearance of Internalized Apotransferrin at the Cell Surface-The rate constant k, cannot be determined directly from a single experiment. However, since we know the constants k l , kZ, and &, we can determine which value of k, gives the best fit to the experimental curve (Fig. 2) showing the fate of '2sI-transferrin during a single cycle of endocytosis. The best value of k, was 0.139 rnin". Thus, the half-time for dissociation of iron and transport of apotrans- ferrin to the surface is 5.0 min, and the mean time for this process is 7.2 min. The theoretical curve for the behavior of transferrin during endocytosis is depicted in Fig. 6. We regard the close agreement with the experimental results as a confir- mation of our model and of the calculated rate constants.

Rate of Dissociation of Apotransferrin from the Surface Transferrin Receptor-In order to measure the rate constant for dissociation of apotransferrin from the recycled receptor (&), we first bind '251-ferrotransferrin to HepGX cells at 4 "c. After washing off excess unbound ligand, iron is released from the bound ferrotransferrin using a combination of low pH and desferrioxamine (3, 14, 15). Under these conditions, the pro- tein moiety remains bound to the cell surface receptor (15). The cells are then incubated a t 37 "C in prewarmed Hanks' solution containing energy inhibitors at pH 7.3. Fig. 7 shows that the half-time for release of the apotransferrin at neutral

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

Transferrin Receptor Recycling 9685

A. Calculated

Surface /

5

z -

0.5 I\ /Med'um

"= 0 10 20

TIME (MIN)

s"L

FIG. 6. Calculated and observed recycling kinetics of the transferrin receptor in HepG2 cells. A, Calculated curue: the differential equation describing our model of receptor internalization and recycling (Fig. 1) was solved for a single endocytic cycle making the following assumptions. At t = 0, (R-Tf.Fe), = 1, ( R - L ) , = 0, R. = 0; klL = 0 (i.e. the absence of free ligand precludes the binding of additional ligand to the cell); k2 = 0.20 rnin", k, = 0.14 rnin", and ko = 2.6 rnin". The number of surface ligands, intracellular ligands, and ligand released to the medium is plotted, normalized to the value of 1.0 a t t = 0. B, Observed curue: the experiment was similar to that described in Fig. 2. At t = 0, 6% of the cell-associated radioactivity could not be removed by pronase. This background value was sub- tracted from all values of cell-associated radioactivity. The amount of cell surface radioactivity was then normalized to a value of 1.0 at t = 0 min.

\ I 5!! W 0 0.1 0 L-"l 10 20 30 40 50

TIME (SEC)

FIG. 7. Dissociation of apotransferrin from the surface re- ceptor of HepG2 cells at pH 7.3. '"I-Ferrotransferrin was bound to cells a t 4 "C in the presence of energy inhibitors and the excess unbound ligand was washed off as described under "Experimental Procedures." 1 ml of 25 mM Na acetate, pH 5, 150 mM NaC1,5 X M desferrioxamine (Desferal Mesylate, Ciba), 5 mM CaC12, and energy inhibitors were added for 5 min a t 4 "C. The medium was aspirated and replaced by prewarmed Hanks' solution containing 128 nM un- labeled transferrin, energy inhibitors, and desferrioxamine. At the indicated time points, the medium was removed, and the cell-associ- ated radioactivity was quantified as described under "Experimental Procedures."

pH is 16 s; thus, & = 2.6 rnin". This rate is 29 times faster than that for dissociation of ferrotransferrin from its receptor ( k 1 ) . This finding was expected, since we had already shown that apotransferrin has a very low affinity for its receptor a t pH 7.3 (15).

Receptor Recycling Time-From Equation 6, we can now calculate T,, the mean time required for a receptor to traverse the entire endocytic cycle, at an external transferrin concen- tration of 6 pg/ml(77 nM): T, = (0.23 min-')" + (0.20 min-')-' + (0.139 min")-' + (2.6 min")-' = 16.9 min.

The Total Number of Functional Transferrin Receptors- From knowledge of the steady state rate of uptake of ligand at 37 "C and the total number of cellular transferrin receptors, we can also determine, completely independently, the mean time required for a receptor to complete a cycle of endocytosis: T, = (total number of receptors: molecules/cell)/(ligand up- take: molecules/cell/min). It is to the first of these numbers that we turn first. Previously, we showed that there are 5 X lo4 transferrin-binding sites/cell surface, with a dissociation constant, kd = 4.4 x lo-' M (6). The assumption that trans- ferrin remains bound to its receptor throughout the endocytic cycle provides us with one direct measure of the total number of transferrin receptors per cell (22). Cells are incubated with a saturating amount of '251-ferrotransferrin at 37 "C. Under these conditions, all functional cellular receptors will partici- pate in the binding and internalization process, and internal- ized surface receptors will be replaced instantaneously by newly recruited intracellular binding sites. When steady state has been achieved, the number of cell-associated ligand mol- ecules should equal the total number of cellular binding sites. This assumes that binding of '251-ferrotransferrin to surface receptors occurs rapidly, so that at all times all the receptors are occupied. We showed that this is indeed the case under the binding conditions employed. Another assumption is that all cell-associated radioactivity represents receptor-bound transferrin, and that transferrin does not dissociate from its receptor and become associated with other intracellular struc- tures, for instance, the lysosomes.

Fig. 8 shows that 1.5 x lo5 '251-ferrotransferrin molecules

1000c

r 0 20 40 60 80 100 200

'z51-TRANSFERRIN (nM)

FIG. 8. Concentration-dependent association of '"1-ferro- transferrin with HepG2 cells. '251-Transferrin was incubated at the indicated concentrations with HepG2 cells for 0.5 h in a medium containing 300 nM 1251-ferrotransferrin and 10" M ferric ammonium citrate to saturate exocytosed apotransferrin. Cell-associated ligand was determined as described under "Experimental Procedures."

a+

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

9686 Transferrin Receptor Recycling

are associated with each cell during steady state ligand uptake at 37 "C (there are 4.0 X lo6 cells/mg of protein). The apparent kd measured at 37 "C was similar to that measured at 4 "C (Refs. 6 and 22; data not shown) indicating that the binding is probably to the same receptor molecule and with the same affinity. The same value (7% higher) was obtained when the binding was continued for an additional 2 h at 4 "C, following the incubation a t 37 "C, in order to saturate any unoccupied surface receptors from which apotransferrin had been disso- ciated at the end of the cycle. As a control to ensure that all of the '"I-labeled cell-associated transferrin is exocytosed normally, cells are incubated with '251-ferrotransferrin for 2 h at 37 "C; unbound ligand is removed, and the cells are incu- bated for 25 min at 37 "C in the presence of 128 nM unlabeled ligand. Cell-associated radioactivity declines to 8% of the initial value, showing that essentially all of the ligand is exocytosed to the medium.

To obtain another estimate of the total number of func- tional transferrin receptors in HepG2 cells, we measured the extent of binding of "'I-transferrin to extracts of detergent- solubilized cells. As can be seen in Table I, the total number of cellular receptors measured in solubilized cells is 1.15 x lo5. Taking into account that about 20% of the transferrin- receptor complex is not precipitated at 32.5% saturation am- monium sulfate (concentration a t which the precipitation was conducted, and at which free transferrin is not precipitated), the total number is 1.44 x lo5, very similar to that obtained from the binding of 12"I-transferrin to cells at 37 "C, 1.5 X le5. We note, however, that the ligand-binding constant in the detergent-solubilized cells is higher, by 1 order of magnitude, than that for ligand binding to cell surface receptors, probably reflecting a lower affinity between the receptor and the ligand under the detergent conditions. As can be further seen from Table I, the number of surface binding sites in single-cell suspension is very similar to that of cells in monolayer, indicating that the additional sites measured in detergent- solubilized cells are all intracellular.

Rate of Transferrin-mediated "Fe Uptake: Independent De- termination of the Cycle Time of Transferrin Receptor-As was shown earlier, the uptake kinetics of iron in a single endocytic cycle is different from that of the transferrin protein moiety. While the iron remains within the cell, the protein diacytoses through the cell, and is exocytosed as apotransfer- rin (14). Thus, the rate of transferrin-mediated iron uptake is a measure of the total rate of transferrin endocytosis. At 4 "C,

TABLE I Binding of '251-ferrotransferrin to intact and detergent-solubilized

cells Conditions 1 and 2 represent 1Z511-transferrin binding to intact

monolayer cells, or to cells released from the dish and kept in single- cell suspension. In condition 3, suspension cells to which '251-trans- ferrin had been bound, were solubilized in detergent. Receptor-bound 1251-transferrin was detected by precipitation with (NH4),S04 as de- scribed under "Experimental Procedures." In 4, suspension cells were disrupted with detergent, '*'I-ferrotransferrin was then added, and receptor-bound transferrin was quantified as described under "Ex- perimental Procedures." All data are derived from Scatchard analysis of concentration-deDendent binding assays.

Conditions No. of bind- ine sites/cell K d

-~ ~

M 1. Binding to intact monolayer cells 4.9 X 104 5.4 X 10-9 2. Binding to intact single cells in sus- 4.5 X lo4 4.1 X lo-'

3. Binding to intact suspension cells, 4.1 X lo4 6.3 X lo-' 4. Binding to detergent-solubilized 11.5 X lo4 7.4 X lo-*

pension

then detergent solubilized

cells

the maximum amount of [5gFe]transferrin that could be as- sociated with the cell was similar to that obtained with lZ5I- ferrotransferrin as expected, since both bind to the same surface receptors. In contrast, at 37 "C, there is a time- dependent uptake of iron, linear for almost 4 h (Fig. 9). The rate of uptake is about 1.9 X IO4 iron atoms/cell/min. Since each transferrin binds two irons, this represents 9.5 X IO3 transferrin molecules that cycle through the cell per min. Taking into account that there are about 1.5 X lo5 transferrin- binding sites/cell, and that the experiment is carried out in the presence of 0.25 mM cycloheximide to block protein syn- thesis (Ref. 5; data not shown) and, therefore, synthesis of new receptor molecules, we calculate that the time required for each receptor molecule to traverse an endocytic cycle is 1.5 X lo5 G9.5 x lo3 = 15.8 min. This is similar to the value (16.9 min) obtained from the sum of the individual rate constants (see "Results" and Table 11).

Effects of Lysosomotropic Agents on the Cycling of Transfer- rin-The low pH in the endocytic vesicle is believed to be an important factor in the dissociation of iron from transferrin and its delivery to the cytoplasm (14, 15, 23). Is the low pH also essential for diacytosis of transferrin, and for recycling of the receptor? To probe these questions, we have utilized several lysosomotropic agents, weak bases known to increase the intralysosomal pH (24). As was reported recently, these agents also raise the pH of the endocytic vesicle (16) and therefore perturb the pH-dependent dissociation of asialoo- rosomucoid from its receptor and its delivery to the lysosome (17). One such agent, NH,Cl, decreases iron uptake into cells (Ref. 14; data not shown); however, the fate of the iron internalized via transferrin was not determined.

We addressed these problems by following the fate of "'I- and [59Fe]transferrin in a single endocytic cycle. Preincuba- tion of the cells a t 4 "C in the presence of up to 50 mM NH4C1, 2 mM chloroquine, 50 mM methylamine, 5 mM amantadine, or 5 mM tributylamine has no effect on the binding of "'I- ferrotransferrin to cell surface receptors. When preincubation with these agents is carried out a t 37 "C for 30 min, the binding is only slightly decreased. With 0.5 mM chloroquine, the inhibition is 32%; with all other agents, the inhibition is much lower (data not shown). At these concentrations, these agents have no effect on the rate of internalization of surface- bound ligand (data not shown).

As is evident from a comparison of Figs. 3 and 10, the same amount of [59Fe]transferrin is internalized in the presence or absence of 20 mM NH4C1. However, while internalized iron in nontreated cells dissociates from transferrin and remains in the cell (Fig. 3), in the presence of NH4C1, the iron is exocytosed into the medium. All other lysosomotropic agents

0 1 2 3 4 TIME (HR)

FIG. 9. Time course of transferrin-mediated "Fe uptake into HepG2 cells. ["FeITransferrin (77 nM) was bound to cells for the indicated times as described under "Experimental Procedures."

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

Transferrin Receptor Recycling 9687

TABLE I1 Parameters of endocytosis of transferrin and asialoglycoprotein

Receptor

Free I

- A.

B.

C.

D.

E.

F.

Parameter Trans- ferrin

Asialo- glyco-

Drotein"

Binding of ligandb k, (mol" min") Mean time (min)'

Dissociation of ligand k-l (min") Mean time (rnin)

k2 (rnin") Mean time

k, (rnin") Mean time

k,, (rnin") Mean time

Measured from the rates of iron or asialoorosomucoid uDtake

Internalization of surface receptor

Return of receptor to surface

Dissociation of apotransferrin

Cycle time (TJ ( m i d

3.02 X lo6 4.3

0.09/0.106d 11.1/9.4

0.20/0.30 5.013.3

0.14 7.19

2.6 0.36

15.8

According to Schwartz et al. (5). Experiments were performed at 50 nM asialoorosomucoid and 77

nM transferrin. Equals ( k1-l) (molar concentration of ligand)"; the mean time

required for a surface receptor to bind a ligand at the concentration employed 50 nM asialoorosomucoid or 77 nM transferrin.

The first measurement was carried out using 1251-ferrotransferrin and the second using [59Fe]transferrin. The former values are used in all subsequent calculations.

2.23 X lo6 8.7

<o.oo

0.46 2.2

0.24 4.2

15.9

15.1

loot

0 1 TIME (HR)

2 3 4

FIG. 10. Effect of NH4Cl on the endocytic cycle of transfer- rin. Experimental details are the same as described in Figs. 2 and 3 except that 20 mM NH,Cl was added to all the solutions. Shown are percentages of radioactivity in cell-associated material that is pronase resistant (i.e. internalized).

tested showed the same effect (data not shown). All of the exocytosed 59Fe radioactivity is present in ferrotransferrin (Fig. 11). '251-Transferrin is exocytosed at the same rate as is [59Fe]transferrin, but a t a rate which is much slower than that in untreated cells (Fig. 2). In NH,Cl-treated cells, a maximum of 64% of surface-bound '251-tran~ferrin can be recovered intracellularly (Fig. IO), compared to 39% in untreated cells (Figs. 2 and 6). This is probably due to an inhibition of the rate of exocytosis of internalized transferrin, with a decrease in k, and a concomitant accumulation of a higher level of transferrin within the cell.

D B Transferrm Iodlne

K

3 50 60 Fract ion Number

FIG. 11. Exocytosisof intact "'1- and ["Fe]transferrin from HepG2 cells in the presence of NH4Cl. The experiment was performed essentially as described in Fig. 10. After 10 min at 37 "C, the medium was removed and quickly replaced with fresh medium also containing 20 mM NH&l and 5 X desferrioxamine. At this time, 87% of the cell-associated radioactivity is resistant to pronase. The incubation was continued for an additional 230 min in the presence of NH4C1 and desferrioxamine. The medium was collected and concentrated under vacuum. It was then chromatographed on a Sephadex G-150 column (1 X 50 cm) equilibrated with 25 mM Na- HEPES, pH 7.5, 150 mM NaC1. Markers are dextran blue (DB) and labeled transferrin. Note that all of the exocytosed 59Fe chromato- graphs with intact transferrin, even though the medium contains sufficient desferrioxamine to prevent binding of any exocytosed free iron to apotransferrin.

We conclude that lysosomotropic agents block the dissocia- tion of iron from transferrin, but do not block the recycling of ferrotransferrin to the surface, presumably still bound to its receptor. Whether NH,Cl slows the rate of diacytosis, i.e. reduces k,, is difficult to determine. Since k - l / ( k - l + k,) = 0.31, only about 30% of ferrotransferrin, recycled to the cell surface bound to its receptor, will be expected to dissociate into the medium; the balance should be re-endocytosed into the cell. Thus, the rate of secretion of i5'Fe]- or lZ5I-transferrin is not a direct estimate of k,, the rate of diacytosis.

DISCUSSION

The human hepatoma line HepG2 provides a well charac- terized continuous cell line with an abundance of transferrin, insulin (6), and asialoglycoprotein receptors (19). An obvious advantage of this monolayer cell line over freshly isolated hepatocytes or cell lines growing in suspension for a study of receptor physiology is that it is a metabolically stable popu- lation of defined cells that can be manipulated rapidly. That these cells have well characterized receptors for a number of polypeptide ligands enables one to compare simultaneously the endocytic rates and pathways taken by different receptors and to study the specificity of the process and its regulation.

Growing HepG2 cells contain about 5 X lo4 surface binding sites for ferrotransferrin and 10 X IO4 intracellular sites. This distribution is different from that of the asialoglycoprotein receptor in growing HepG2 cells, where 87% is on the cell surface ( 5 ) . This distribution is, however, similar to that found for transferrin receptors (albeit indirectly) in teratocarcinoma cells (13) and in HeLa cells (22). I t is also similar to the distribution described for the mannose receptor in alveolar macrophages (8). We emphasize that the distribution of re- ceptor in growing cells and during ligand uptake is a reflection

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

9688 Transferrin Receptor Recycling

of many aspects of cell growth and physiology. HepG2 cells are grown and maintained in culture media which contain 10% fetal calf serum. Serum contains transferrin which can probably affect the number (22) and distribution (14) of transferrin receptors. In addition these cells actively synthe- size human serum proteins (18) (including transferrin'), which may alter the number or localization of transferrin receptors.

In HepG2 cells, uptake of transferrin occurs for a t least 4 h, at a steadystate rate of 9.5 X io3 protein molecules/cell/ min, independent of new receptor synthesis. We conclude that, during this period of time, each receptor must be utilized about 15 times, with a cycling time of about 15.8 min.

The steady state uptake of ligand is adequately described by the kinetic model shown in Fig. 1 and by the values in Table 11. As can be seen in Table 11, many of the values have been determined by more than one independent type of ex- periment. The close agreement between the measurements attests to the general validity of our methods. For instance, the half-time for internalization (3.5 min) could be determined from the rate of loss of surface-bound 1251-ferrotransferrin, in the presence and in the absence of energy inhibitors (Fig. 4). It could also be detemined by measuring the time necessary for internalization of 50% of [59Fe]transferrin bound to the cell surface (2.3 min; Fig. 3). k l , the rate constant for disso- ciation of transferrin from its surface receptor, was also determined by measuring either the dissociation of prebound '"I-ferrotransferrin in the presence of energy inhibitors (7.7 min, Fig. 4) or directly, by dissociation into the medium of prebound [59Fe]transferrin (6.5 min, Fig. 3). An additional correlation comes from the measurement of the ratio of the internalized transferrin molecules compared to those which are dissociated directly from cell surface receptors prior to internalization. For '*"-transferrin, the calculated ratio is k2/ ( k 1 + k,) = 0.69 and the measured value for [59Fe]transferrin is 0.63 (Fig. 3).

We emphasize that using two different labeled ligands is not simply a change in the radioisotope, but rather measuring two different pathways. Because the transferrin protein moiety recycles, the measurements of ligand internalization are therefore somewhat indirect and require the use of energy inhibitors. Since the iron moiety stays in the cells, many rate constants can be measured in a more direct way. As a final example of the validity of our analysis, the cycling time of a receptor as calculated from the total number of cellular bind- ing sites and the iron uptake rate (15.8 min, Table 11) is in very good agreement with the time calculated from the sum of the mean times of each separate step of the cycle (16.9 min; see "Results" and Table 11). We note, however, that some of the important assumptions upon which the model is founded, namely steady state conditions and unaltered number of functional receptors during the observation period, are not absolute.

It is of interest to compare the kinetic parameters for endocytosis of transferrin to those for asialoglycoprotein in the same cell line (Table 11). The various parameters of the endocytic cycle are similar although some differences should be mentioned. The rate constants for ligand binding, kl , for the asialoglycoprotein and the transferrin receptors differ by 50%. The mean time required for a surface receptor to bind to an asialoglycoprotein, 8.0 min, was obtained at a ligand concentration of 50 nM (50% saturation of the receptor) while that for transferrin, 4.3 min, was obtained at 77 nM (>go% saturation). Obviously, at greater ligand concentration, the

H. F. Lodish, N. H. Kong, M. Snider, and G. J. A. M. Strous, manuscript in preparation.

time required for ligand binding will decrease, until a point is reached where binding occurs in an infinitely short time. The klL values, however, which take into account the concentra- tion, are very similar (3.02 X lo6 for transferrin and 2.23 X lo6 for asialoorosomucoid, respectively; Table 11).

The rate constants for internalization of the two ligands are slightly different; surface-bound asialoglycoprotein is in- ternalized 1.5-2.5 times faster than is transferrin bound to surface receptors. The times are within the same range and might be due to experimental variation. The rate constants for recycling of the two internalized receptors to the cell surface, k,, are also similar; a mean time of 4-7 min is required for these processes. Whether these two receptors (and possibly others) share the same endocytic pathway cannot, of course, be determined from this type of kinetic study. The epidermal growth factor receptor and the low density lipoprotein recep- tor have been shown to share the same coated pits and vesicles during receptor-mediated endocytosis (25).

We note that the cycling time of both the transferrin and asialoglycoprotein receptors is rather short, around 15 min. The time necessary for internalization of the receptor-ligand complex into the low pH endocytic vesicle and the dissociation of the ligand or the coligand (in the case of transferrin) is even shorter. Dissociation of asialoglycoprotein from its re- ceptor probably occurs in a low pH endocytic vesicle (2), designated CURL (compartment of uncoupling of receptor and ligand) by Geuze et al. (1). However, degradation of many ligands in lysosomes starts only after a lag period of about 30 min (5,26). This delay is probably due to slow delivery of the ligand to the lysosome or to a delayed action of the lysosomal proteases, but not to any step in which the receptor is in- volved.

Lysosomotropic agents such as NH4C1 do not inhibit bind- ing of transferrin to cell surface receptors, nor do they inhibit internalization of the receptor-ligand complex (Ref. 14; data not shown). They do have two marked effects on the trans- ferrin endocytic cycle; they slow the exocytosis of the trans- ferrin polypeptide, although all endocytosed transferrin is eventually secreted. They also block the retention of iron in the cell; all of the endocytosed Fe is secreted into the medium as diferric transferrin. Exocytosis of [59Fe]transferrin in the presence of chloroquine was also reported by Octave et al. (27). However, the effect could be detected only after prein- cubation of the cells in the presence of the drug for 16 h a t 37 "C, and could not be shown with other lysosomotropic agents. Furthermore, the analysis was not done in a single endocytic cycle, which makes their results even more difficult to interpret.

The principal effect of NH4Cl is the inhibition of dissocia- tion of iron from the endocytosed transferrin-receptor com- plex, presumably because it raises the pH of the endocytic vesicle. Iron is dissociated from the transferrin-receptor com- plex only at pH values less than 6.0. NH4C1 reduces signifi- cantly the rate, although not the extent, of diacytosis of transferrin polypeptide into the medium. There are several possible explanations for this. 1) Ferrotransferrin that recy- cles to the cell surface bound to its receptor may be subject to rapid endocytosis; by contrast, all apotransferrin that is recycled to the cell surface is rapidly released into the medium, since apotransferrin dissociates much faster than transferrin from the receptor at neutral pH (see "Results" and Ref. 16). 2) The ferrotransferrin-receptor complex might recycle to the surface slower than the apotransferrin-receptor complex. 3) The high pH might decrease the rate of recycling regardless of whether or not iron is dissociated from the apoligand. In the case of the asialoglycoprotein (28), mannose (291, and mannose 6-phosphate receptors (7), it was shown that lyse-

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

Transferrin Receptor Recycling 9689

somotropic agents reduce the amount of detectable ligand 11. Marshall, S., Green, A., and Olefsky, J. M. (1981) J. Bwl. Chem. binding at the cell surface. Although no definitive mechanism 256,11464-11470 is yet available, it is assumed that these agents also decrease 12. Jan& J. H., and Katz, J. H. (1963) J. C h . Invest. 42,314-326 the rate of recycling of these receptors. any case, these 13. Karin, M., and Mintz, B. (1981) J. Bid. Chem. 256,3245-3252

experiments show that the low pH of the endocytic vesicle is 14. Klausner, R. D., van Renswoude, J., Ashwell, G., Kempf, C.,

essential for dissociation of iron from the transferrin-receptor Schechter, A. N., Dean, A., and Bridges, K. R. (1983) J. Biol. Chem. 258,4715-4724

complex, but is not essential for the recvcling of the transfer- 15. Dautm-Varsat. A.. Ciechanover. A.. and Lodish. H. F. (1983) - . rin polypeptide back to the cell surface.

Acknowledgments-We are grateful to Dr. Richard D. Klausner, 17, 16.

for making his results available to us prior to publication. We thank Miriam Boucher for her skillful help and patience in the preparation 18, of the manuscript.

"

1.

2. 3.

4.

5.

6.

7.

8.

9.

10.

REFERENCES Geuze, H. J., Slot, J. W., Strous, G. J. A. M., Lodish, H. F., and

Tycko, B., and Maxfield, F. R. (1982) Cell 28,643-651 Klausner, R. D., Ashwell, G., van-Renswoude, J., Harford, J. B.,

and Bridges, K. R. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 2263-2266

Tanabe, T., Pricer, W. E., Jr., and Ashwell, G . (1979) J. Bwl. Chem. 254, 1038-1043

Schwartz, A. L., Fridovich, S. E., and Lodish, H. F. (1982) J. Biol. Chem. 257,4230-4237

Ciechanover, A., Schwartz, A. L., and Lodish, H. F. (1983) Cell

Gonzalez-Noriega, A,, Grubb, J. H., Talkad, V., and Sly, W. S. (1980) J. Cell Biol. 85, 839-852

Stahl, P., Schlesinger, P. H., Sigardson, E., Rodman, J. S., and Lee, Y. C. (1980) Cell 19, 207-215

Brown, M. S., Anderson, R. G. W., Basu, S. K., and Goldstein, J. L. (1982) Cold Spring Harbor Symp. Quunt. Biol. 46,713-721

Van Leuven, F., Cassiman, J-J., and Van Den Bergh, H. (1980) Cell 20, 37-43

Schwartz, A. L. (1983) Cell 32, 277-287

32,267-275

19.

20.

21. 22.

23.

24.

25.

26.

27.

28.

29.

Pro;. Natl. Acad: Sci. U. S. A. 80,'2258-2262 '

, ,

Maxfield, F. R. (1982) J. Cell Biol. 95,676-681 Harford, J., Bridges, K., Ashwell, G., and Klausner, R. D. (1983)

Knowles, B. B., Howe, C. C., and Aden, D. P. (1980) Science

Schwartz, A. L., Fridovich, S. E., Knowles, B. B., and Lodish, H.

Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J .

Wiley, H. S., and Cunningham, D. D. (1981) Cell 25,433-440 Ward, J. H., Kushner, J . P., and Kaplan, J. (1982) J. Biol. Chem.

van Renswoude, J., Bridges, K. R., Harford, J . B., and Klausner,

Ohkuma, S., and Poole, B. (1978) Proc. Natl. Acad. Sci. U. S. A.

Carpentier, J. L., Gorden, P., Anderson, R. G. W., Goldstein, J. L., Brown, M. S., Cohen, S., and Orci, L. (1982) J. Cell Biol. 95,73-77

Goldstein, J . L., and Brown, M. S. (1974) J. Biol. Chem. 249,

Octave, J-N., Schnieder, Y-J., Hoffman, P., Trouet, A., and

Tolleshaug, H., and Berg, T. (1979) Biochem. Pharmacol. 28,

Tietze, C., Schlesinger, P., and Stahl, P. (1980) Biochem. Biophys.

J. Biol. Chem. 258,3191-3197

(Wash. D. C.) 209,497-499

F. (1981) J. Biol. Chem. 256,8878-8881

(1951) J. Biol. Chem. 193, 265-275

257,10317-10323

R. D. (1982) Proc. Natl. Acad. Sci. L'. S. A. 79, 6186-6190

75,3327-3331

5153-5162

Crichton, R. R. (1982) Eur. J. Biochem. 123,235-240

2919-2922

Res. Commun. 93, 1-8

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from

A Ciechanover, A L Schwartz, A Dautry-Varsat and H F Lodishreceptor in a human hepatoma cell line. Effect of lysosomotropic agents.

Kinetics of internalization and recycling of transferrin and the transferrin

1983, 258:9681-9689.J. Biol. Chem. 

  http://www.jbc.org/content/258/16/9681Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/258/16/9681.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on July 1, 2020http://w

ww

.jbc.org/D

ownloaded from