Importance of Ligand-induced Conformational Changes in Hemopexin

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 11. Iasue of April 15, pp. 5224-5229,1988 Printed in U.S.A.

Importance of Ligand-induced ConformationalChanges in Hemopexin for Receptor-mediated Heme Transport* (Received for publication,May 20,1987)

Ann Smith*, Fred M. Tatum, Peter Muster, Mary K.BurchQ,and William T. Morgan From the Demrtment of Biochemistry and Molecular Biology, Louisiana State University Medical Center, New Orleans, Lo~isianu~ 70112

Hemopexin alters conformation upon binding heme The transport of heme' to the liver by hemopexin is meas shown by circular dichroism (CD), but hemopexin diated by a specific receptor in the plasma membrane of the binds the heme analog, iron-meso-tetra-(4-sulfonato- liver parenchymal cells (1-3). Upon binding heme, hemopexin phenyl)-porphine (FeTPPS),without undergoing con- undergoes a change in conformation which was proposed to comitant changes in its CD spectrum. Moreover, Fe- be important for increasing the affinity of the heme-hemoTPPS, unlikeheme, does not increase thecompactness pexin complex for the hemopexin receptor (4). These changes of the heme-binding domain (I) of hemopexin shown by in conformation, ascribed in part to tryptophan residues of an increased sedimentation rate in sucrose gradients. hemopexin (4), have been demonstrated using absorbance (5) On the other hand, like heme, FeTPPS forms a bis- and circular dichroism (CD) spectroscopy (4, 6). In addition, histidyl coordination complex with hemopexin and binding of heme affords protection to hemopexin from proupon binding protects hemopexin from cleavage by teolysis (7)which produces two active domains (8) linked by plasmin. Competitive inhibition and saturation studies a hinge region (9). The evidence to date suggests that the interaction of heme-hemopexin with its receptor results in demonstrate that FeTPPS-hemopexin bindstothe hemopexin receptor on mouse hepatoma cells but with the removal of the heme from hemopexin and the return of a lower affinity (& 125 nM) more characteristic of intact apo-hemopexin to the circulation (3, 10). The heme is apo-hemopexin than heme-hemopexin ( K d 65 nM). This transferred to an intrinsic heme-binding membrane protein provides evidence that conformational changes pro- (MHBP) which is considered to transport heme into the cell duced in hemopexin upon binding heme, but not upon (10). binding FeTPPS, are important for increasing the af- Nonetheless, major questions remain as to theexact nature finity of hemopexin for its receptor. The amount of of the conformational changes in hemopexin induced by its cell-associated radiolabel from "FeTPPS-hemopexin ligand, heme, and the role of these changes in recognition of increases linearly for up to 90 min but at a rate only heme-hemopexin by its receptor. To obtain further informaabout a third of that of the mesoheme-complex. As tion on the details of the interaction of the heme-hemopexin expected from the recyclingof hemopexin, more iron- complex with its receptor, conformation- and ligand-associtetrapyrrole than protein is associated with the Hepa ated determinants of hemopexin-mediated heme transport cells, but the ratio of "'Fe-ligand to 'Z61-hemopexinis were addressed using the heme analog, iron-meso-tetra-(4only 2:l for FeTPPS-hemopexin compared to 4:l for sulfonatopheny1)-porphine,also commonly called iron-tetracompound is mesoheme complexes. [5"Fe]Mesohemewas associated phenylporphine sulfonate (FeTPPS).This bound tightly by hemopexin (1:l stoichiometry, Kd below 10 at 5 min with lower densityfractionscontaining plasma membranes and at 30 min with fractions con- nM, Ref. 11, comparable to the Kd of the heme-hemopexin taining higher density intracellular compartments. In complex; Ref.12), but it is more symmetric and water-soluble contrast, 56FeTPPSwas found associated with plasma than the amphipathic, naturally occurring hememolecule. membrane fractions at both times and was not trans- Importantly, FeTPPS is shown here to produce only some of portedintothe cell. Although FeTPPS-hemopexin the changes in conformation caused by the binding of heme binds to the receptor, subsequent events of heme trans- to hemopexin, enabling a clearer definition of these changes port are impaired. The results indicate that upon bind- and of their importance in the biological function of hemoing heme at least three types of conformational changes pexin. occur in hemopexin which have important roles in MATERIALS AND METHODS receptor recognition and that the natureof the ligand influences subsequent heme transport. Hemopexin was isolated from rabbit serum, and its purity (greater than 95%) checked as previouslydescribed (13). Sodium dodecyl sulfate-polyacrylamide slab gel electrophoresiswas carried out using a 4-20% acrylamide gradient. The two domains of rabbit hemopexin wereobtained by ion-exchange chromatography after digestion of apo-hemopexin with plasmin using published methods (7).Proteins

* Supported inpart by United States Public Health Service Grant The abbreviations used are: heme, iron-protoporphyrin IX, meDK 27237 (to W. T. M.) and by a fellowship fromthe Swiss National Foundation (to P. M.). The costs of publication of this article were soheme, iron-mesoporphyrinIX; FeTPPS, iron-meso-tetra-(4-sulfondefrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 To whom correspondence should be addressed. Present address: Rohmand Haas Co., 727 Norristown Rd., Spring House, PA 19477.

atopheny1)-porphine,also commonly callediron-tetraphenylporphine sulfonate; TPPS, theiron-free form of FeTPPS; MHBP, membrane heme-binding protein component of hemopexin transport system; DMEM, Dulbecco's minimal essentialmedium; DEP, diethylpyrocarbonate; TES, 10 mM triethanolamine, 1 mM EDTA, 0.25 M sucrose, pH 7.5;HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.

5224

Determinants of Hemopexin Transport were labeled with lZ5I(ICN, Irvine, CA) using Iodobeads (Pierce Chemical Co.) according tothe manufacturer's instructions, and unincorporated "'I removed by passage over Sephadex G-25(Sigma) or by extensive dialysis. Iodinated proteins were routinely more than 95% precipitable by 12.5% final volume trichloroacetic acid and had a specific activity of 4000-5000 dpm/pmol. Mesoheme (iron-mesoporphyrin IX) from Porphyrin Products (Logan, UT) was used in place of heme (iron-protoporphyrin IX) since it is more stable and mesoheme-hemopexin complexes are chemically and biologically equivalent to heme-hemopexin (1-3). Mesoporwere phyrin, FeTPPS, and meso-tetra-(4-sulfonatophenyl)-porphine also from Porphyrin Products. Porphyrins were labeled with 55Fe(Du Pont-New England Nuclear) by refluxing in dimethylformamide (14) and freed of unincorporated iron and metal-free porphyrin by washing with acid (1). Complexes of hemopexin or of heme-binding domain I with mesoheme and FeTPPSwere prepared by mixing 1 eq of tetrapyrrole with 1 eq of protein. Unbound tetrapyrrole was removed by passage over DEAE-cellulose (DE52, Whatman) (1) or by extensive dialysis at 4 "C. Concentrations were determined spectrophotometrically using extinction coefficients (A M" cm-') of 1.1 X 105 at 280 nm for apohemopexin (15),of 1.3 X IO5 a t 405 nm for mesoheme-hemopexin, of 1.2 X lo5 at 403 nm for mesoheme-domain I (7),of 1.7 X lo5 at 394 nm for mesoheme in dimethyl sulfoxide (16),and of 1.5 X 10' at 392 nm for FeTPPS in 0.1 M NaCl, pH 3.3 (11). Absorbance measurements were recorded using a Cary 219 spectrophotometer and CD measurements with a Jasco 500C spectropolarimeter. Heme-hemopexin is in the ferric state under aerobic conditions (6),and reduced spectra were obtained by adding a few crystals of sodium dithionite. Modification reactions using diethylpyrocarbonate (DEP) were carried out for 1h on ice by adding the indicated amount of DEP to apo- or mesoheme-domain I (at a concentration of 4.6 p M ) in 0.1 M sodium phosphate, pH 6.3. The number of DEP-modified histidine residues was calculated using a millimolar difference extinction coefficient of 3.2 a t 240 nM (17).Solutions of DEP were prepared in absolute ethanol. Effects of modification on heme binding were assessed as previously described (18) by the loss of absorbance in the characteristic Soret band of the mesoheme-domain complex a t 405 nm. Sucrose gradients (5-20% w/v, 5.8 ml) were run in aBeckman VTi 80 rotor a t 4 "C and 463,000X g (80,000 rpm) to a preset cumulative centrifugal effect (u't) of 3.04 X lo4 rad' s-I. After mixing domain I with ligand (2.5p M final concentration of each), the mixtures were incubated on ice for 15 min before 200 plwas layered onto the gradient and centrifuged. After the run, 10-drop fractions were collected from the bottom of the centrifuge tube and their radioactivity measured. In a parallel run labeled chymotrypsinogen (2.6S), ovalbumin (3.5 s),aldolase (7.9 s),and @amylase (9.4S) were used as standards to determine the sedimentation coefficient of the protein. Minimal deviation hepatoma cells (mouse Hepa cells) from mouse solid tumor line BW 7756 were kindly provided by Dr. BarryLedford, University of South Carolina, and grown in Dulbecco'smodified Eagle's medium (DMEM) containing 2% fetal calf serum and gentamycin (50 pglml). For all studies the Hepa cells were maintained in the log phase of growth. This cell line has been shown to synthesize hemopexin and to possess all the cellular components needed for specific, hemopexin-mediated cell uptake of heme? Binding of '''Ilabeled protein was measured intriplicate wells by adding 1-ml aliquots of Hepes-buffered DMEM, pH 7.4,containing the protein to 1.0-1.5 X lo6 cells in one well of a 6-well tissue culture dish. At the indicated time, the medium was aspirated from the well and thecells washed with 3 X 2 ml of cold 10 mM sodium phosphate buffer, pH 7.4,containing 0.15 M NaCl. Then NaOH (0.1 M, 2 ml) was added to dissolve the cells, followed by0.23 ml of 50% acetic acid to neutralize the dissolved cells, and 2.0 ml counted in a Beckman 7500 gamma spectrometer. At least 3 wellswere used for each determination. Hemopexin-mediated uptake of 55Fe-labeledmesoheme or FeTPPS was measured in 2-ml aliquots in a similar manner in 13 ml of liquid scintillation fluid (Betablend, Westchem Scientific, San Diego, CA) using a Beckman 9800 scintillation counter. Specific binding or uptake is defined as thedifference in binding or uptake inthe presence and absence of excess unlabeled mesoheme-hemopexin added to the cells before the labeled complex. Additional details are given in the legends to figures and tables. Cell counts were made on at least 2 wells before each experiment by releasing the cells with trypsin and A. Smith andB. E. Ledford, submitted for publication.

Functions

300

400

450

5225

500

550

600

WAVELENGTH (nm) FIG. 1. Absorbance spectra of FeTPPS-hemopexin. Absorbance spectra of equimolar FeTPPS-hemopexin were recorded in 10 mM sodium phosphate buffer, pH 7.4,containing 0.15 M NaCl before (solid line)and after (dashed line) reduction with dithionite. The concentration of the complex was 6.5 pM. Also shown is FeTPPS at the same concentration in the same buffer containing 0.1 M imidazole (dash-dotline,before reduction; dotted line, after reduction). counting the cells in a hemocytometer. Protein was determined on aliquots of the neutralized cell extracts using the Pierce BCA protein assay system according to the manufacturer's directions with bovine serum albumin as a standard. Percoll subcellular fractionation was performed essentially according to Dickson et al. (19).Hepa cells in 150 cm' flasks were incubated at 37 "C with 1 p~ 55FeTPPS-or [55Fe]mesoheme-hemopexin.After 5 min, the cells were washed to remove labeled complex, one set was allowed to incubate for an additional 25 min, and one set was washed with cold phosphate-buffered saline. The cells were detached from the flask by scraping with a rubber policeman and homogenized on ice using a tight-fitting Dounce homogenizer in 2 ml of TES buffer. Cell disruption was monitored by staining with trypan blue. The homogenate was centrifuged at 3000 X g for 10 min at 4 "C, and 0.5 ml of the supernatant was layered over 9 mlof iso-osmotic Percoll (Pharmacia LKB Biotechnologies Inc.). Following centrifugation at 40,000X g for 1 h, the gradients were collected from the top with a Haake-Buchler apparatus and assayed for distribution of heme-label and in parallel tubes for density with gradient marker beads (Pharmacia LKB Biotechnologies Inc.) and organelle marker enzymes (19). The second set of flasks was treated in the same way after 25 min of additional incubation. RESULTS AND DISCUSSION

Hemopexin has been shown to bind a variety of heme analogs ( 4 , l l ) . One of these, FeTPPS,is of particular interest. FeTPPS is more symmetrical and more water soluble than heme, and itis bound tightly by hemopexin (1:l stoichiometry, K d below 10 nM, Ref. 11, comparable to the K d of the hemehemopexin complex, Ref. 12).3 However, to allow comparison of this analog with heme, it was necessary first to determine whether hemopexin binds FeTPPS as it does heme in a bishistidyl coordination complex (4, 6). Spectral and chemical modification studies were carried out using both hemopexin and the heme-binding domain I of hemopexin (7), shown elsewhere to bind heme in an analogous manner to hemopexin4but which affords a simpler system for study. As shown in Fig. 1,binding of FeTPPS by rabbit hemopexin shifts the Soret maximum of the ligand to 419 nm (in contrast to the previous report of 423 nm, Ref. ll), and both the oxidized and reduced absorbance spectra of FeTPPS-hemoA. Smith andW. T. Morgan, unpublished observations. W. T. Morgan, P. Muster, F. Tatum, J. McConnel, T. P. Conway, P. Hensley and A. Smith, submitted for publication.

Determinants of Hemopexin Transport Functions

5226

pexin resemble those of imidazole-FeTPPS indicating that it is a low-spin, bis-histidyl heme coordination complex, It is of note that there is no absorbance band at 620 nm characteristic of high-spin proteins (17). In order to obtain additional evidence that 2 histidine residues are the axial ligands of the iron of the bound heme analog, domain I and FeTPPS-domain I were treated with the selective histidine modifying reagent DEP (Table I). Modification of the 5 histidine residues of apo-domain I abolished its ability to form a coordination complex with FeTPPS.Incontrast, only 3 of 5 histidine residues of FeTPPS-domain I couldbe converted to N ethoxyformyl-histidine (Table I) showing protection by the bound FeTPPS of two histidines, and the treatment had no effect on the absorbance spectrum of the complex (not shown). These results confirm that FeTPPS forms a bis-histidyl coordination complex with hemopexin as does heme. However, formation of the FeTPPS-hemopexin complex with a bis-histidyl coordination complex with the central iron atom of the tetrapyrrole does not induce the same changes in the conformation of hemopexin which occur upon binding heme (4,5).This is illustrated by both the ultraviolet absorbance and CD spectra of the twocomplexes. The distinct sharpening of the shoulder at 290 nm in the absorbance spectrum of the protein seen upon binding heme ( 5 ) does not occur (Fig. 2). The different effects of these two ligands on the protein are even more clearly seen by CD spectroscopy. The characteristic increase in ellipticity at 231 nm which occurs when hemopexin binds heme (4, 18) is not produced by FeTPPS (Fig. 3). Both the sharpening at 290 nm and the increase in ellipticity at 231 nm have been ascribed to tryptophan-associated conformational changes (4), so that in some respects the conformation of FeTPPS-hemopexin appears to more closely resemble that of apo-hemopexin than of heme-hemopexin. Additional evidence that FeTPPS does not affect the conformation of hemopexin in thesame manner as heme is provided by the results of sucrose gradient sedimentation experiments. As shown in Fig. 4, the clear increase in the sedimentation rate of domain I of hemopexin upon binding heme, shown elsewhere to reflect a change in shape of domain I,‘ is not caused by the binding of FeTPPS. Nonetheless, hemopexin does change conformation upon binding FeTPPS. This is shown by the protection afforded by both FeTPPS andmesoheme against cleavage of the hinge region by plasmin (Fig. 5 , lanes d-e, h-i). This conformational change conferring resistance to proteolysis requires coordination of the iron atom in the center of the tetrapyrrole ring by histidine residues, since binding of iron-free mesoporphyrin (Kd near 1 PM, Ref. 18) or of tetra-phenylporphine sulfonate (Kd near 1 PM, Ref. 11) do not afford protection (Fig. 5 , lanes f-g and j-k, respectively). In viewof these observations, the heme analog FeTPPS wasused to examine furtherthe role of conformational TABLEI

1.2

0.8

0.4

250

I

I

1

300

350

400

450

WAVELENGTH (nm)

FIG. 2. Ultraviolet absorbance spectra of FeTPPS-hemopexin and mesoheme-hemopexin. Absorbance spectra of equimolar FeTPPS-hemopexin (solidline) and mesoheme-hemopexin (dashed line) were recorded in 10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaCl. The concentrations of the complexes ; line) in were 13 &M. Also shown is FeTPPS alone (11 p ~ dash-dot the same buffer. The inset depicts a magnified viewof the ligandprotein spectra near 290 nm together with that of apo-hemopexin (dotted line)for comparison.

0 c-

X

E’ 0

I-

a J

J

w

I

220

I

230

240

250

WAVELENGTH (nm)

apo-domain I

FeTPPS-domain I

FIG. 3. Circular dichroism spectra of FeTPPS-hemopexin and mesoheme-hemopexin. Circular dichroism spectra of apohemopexin (circles),FeTPPS-hemopexin (triangles),and mesohemehemopexin (closed circles) were recorded a t 25 “C in phosphatebuffered saline, pH 7.4. The concentrations of the protein and the equimolar complexes were 1.6 WM.Data are presented as molar ellipticity, [e] deg .cm2.dmol”.

0 2.7 4.9 5.3 5.3

0 1.1 2.7 3.4 3.4

changes in hemopexin in effecting the interaction of hemopexin with its receptor. The specific binding of mesohemehemopexin and of FeTPPS-hemopexin tothe hemopexin receptor and subsequent release of heme for transport were examined using mouse Hepa cells in vitro. This minimum-

Histidine modification of domain Z with diethylpyrocarbonate Modification reactions were carried out as described under “Materials and Methods” by adding diethylpyrocarbonate (DEP) toapoand FeTPPS-domain I in 0.1 M sodium phosphate, pH 6.3. No. of modified histidine residues

DEP

1.6

M x 10“

0 1.7 5.2 10.4 15.5

Determinants of Hemope,xin Transport Functions M I

100

-

0

80

X

z

2 60 0

40 \

0

0

5

10

15 2520

30

FRACTION NUMBER FIG. 4. Sucrose gradient centrifugation of the complexesof mesoheme and FeTPPS with domain I. 1251-DomainI was centrifuged in 5-20% sucrose gradients as described under “Materials and Methods.” Shown are: apo-domain I (circles), mesoheme-domain I (squares), and FeTPPS-domain I (triangles). Fraction 1 is the

5227

TABLE I1 Competitive inhibition of the interaction of mesoheme- with m u s e Hepu cells The ability of the agents shown to inhibit the binding of mesoheme‘251-hemopexinto mouse Hepa cells was assessed by incubating the competing agent with the cells a t 2.5 p~ for 15 min at 37 “C as described under “Materials and Methods.” The labeled heme-hemopexin complex (50 nM) was then added and binding assessed 15 min later. Under these conditions specific binding was -80% of the total, i.e. the amount of mes~heme-’~~I-hemopexin bound to the cells was decreased to -20% by incubation with nonradioactive mesohemehemopexin. Inhibition of hemopexin-mediated heme transport by mouse hepatoma cells was assessed by incubating the agent (2.5p ~ ) with the cells a t 37 “C for 30 min before adding 250 nM [&Fe] mesoheme-hemopexin.Heme accumulation, linear for up to 2 h, was measured 60 min later. Data presented are the mean f S.D.of a representative experiment, triplicate wells per experiment, and each experiment was repeated a t least twice. Specific binding and heme uptake were 0.24 and 4.5 pmol/mg of protein, respectively, in this exDeriment. uptake Heme Receptor binding Agent Mesoheme-hemopexin FeTPPS-hemopexin Apo-hemopexin Mesoheme FeTPPS

heaviest and fraction 30 the lightest fraction.

%CORtfd

% CORtlVl

21 f 2 59 f 10 48 f 4 86 2 6 90 f 14

40 f 1 75 f 2

78 f 5 106 f 10 111 f 11

ABCDEFGHI J K L 30

94

x 25

Y

-43

L0 20 0

o 15

z 3

-3 1 ‘23 ‘14.4 FIG.5. Protection of hemopexin from proteolysis by bound ligands. Hemopexin was treated with plasmin in the presence and absence of ligands, electrophoresed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis and stained with CoomassieBlue. Shown are: apo-hemopexin after 0 ( h n e b) and 1 h (lane c) exposure to plasmin a t a 150 ratio in phosphate-buffered saline at 25 “C. Lunes d and e contained mesoheme-hemopexin after exposure to plasmin for 0 and 2.5 h, respectively; lanes f and g contained mesoporphyrinhemopexin, lanes h and i contained FeTPPS-hemopexin, and lanes j and k contained TPPS-hemopexin after exposure for 0 and 2.5 h, respectively. The outer lanes (a and I ) contained molecular weight standards as indicated. Approximately 20 pg of protein were applied to each lane.

0

m 10

\

- 5 0 1

/ [COUPLEX]. .u-1

TIME. minuter

FIG.6. Time course and saturationof hemopexin-mediated uptake of FeTPPS and mesoheme by mouse Hepa cells.Panel

A , double-reciprocal plot of cell-associated 1251-hemopexin fromincreasing concentrations of FeTPPS-’251-hemopexin(squares) and mes~heme-’~~I-hemopexin (circles) incubated with mouse Hepa cells for 15 min a t 37 “C. Numerical analysis of the untransformed data using a nonlinear regression analysis yielded apparent K d values of 125 a i d 65 nM and maximum binding of 0.5 and 1 pmol/mg protein for the FeTPPS- andmesoheme-hemopexincomplexes, respectively. Panel B, “FeTPPS-hemopexin (squares) and [55Fe]mesoheme-hemopexin (circles) were incubated with mouse Hepa cells a t 37 “C in Hepes-buffered DMEM, pH 7.4. Cells werein thelog phase of growth, and each well contained -1.5 X lo6 cells. At the indicated times the deviation hepatoma cell line synthesizes and secretes hemo- medium was removed,the cells rinsed with ice-cold buffer and solupexin, expressesboth the hemopexin receptor and the MHBP bilized for determination of cell-associated 65Fe.

(3), and carries out intracellular transport of heme? To demonstrate that FeTPPS-hemopexin binds to the hemopexin receptor on Hepa cells, the competitive inhibition studies summarized in Table I1 were carried out. FeTPPShemopexin proved to be an effective inhibitor of both the binding of lZ5I-labeledmesoheme-hemopexin and of heme uptake from [65Fe]mesoheme-hemopexin.As judged by its inhibitory activity, FeTPPS-hemopexin, like apo-hemopexin (2), binds to the receptor but with a lower apparent affinity (apparent K d of 125 nM) than the heme-protein (apparent & of 65 nM, seebelow).Adding either unbound FeTPPS or unbound mesohemedoes not significantly decreasehemopexin binding to its receptor or hemopexin-mediated heme uptake, demonstrating that the observed inhibitory effects require the specific hemopexin-receptor interaction. To further define differences between the interaction of

FeTPPS-hemopexin and mesoheme-hemopexincomplexes with Hepa cells, the interaction of the complexes with the cells was examined more closely. The binding of FeTPPS’251-hemopexin, like that of mesoheme-hemopexin, withthese cells at 37 “C is saturable as demonstrated by the doublereciprocal plot in Fig. 6A. Analysis of the untransformed data by computer-aided curvefitting yielded an apparent K d of 125 nM for the FeTPPS-’251-hemopexin-receptorinteraction and an extrapolated maximum bindingof 0.5 pmol/mg protein. In contrast, mesoheme-’Z51-hemopexin has an apparent K d near 65 nM and 1 pmol/mg protein is bound at saturation. The amount of cell-associated ‘251-hemopexin isabout one-third lower forFeTPPS thanfor mesoheme-complexes over a range of concentrations (Fig. 7). Although this is consistent with the lower affinity of the FeTPPS-hemopexincomplex forthe

Determinants of Hemopexin Transport Functions

5228 I U

c

2.5

0 E

2.0

W f

1.5

z o_ I-

n

0

L

\ 0

5

0.4

\

1.0

z -I

0

m

d

0.6

0.5

B

o.2 0.0

0.0

0

5

10

15

20

25

FRACTION NUMBER FIG. 8. Isopycnic density gradient fractionation of mouse FIG. 7. Effect of concentration on cell-associated "6FeTPPS, Hepa cells after hemopexin-mediateduptake of E6FeTPPSand ['"Fe]mesoheme, and '"'I-Hemopexin. Mouse Hepa cells were [mFe]mesoheme.Mouse Hepa cells were incubated with 55FeTPPSincubated with the indicated concentrations of FeTPPS-'251-hemo- hemopexin or with [55Fe]mesoheme-hemopexin at 37 "C. The complex pexin (right jine hatch, ///), 65FeTPPS-hemopexin (lejt f i n e hutch, concentration was 0.25 pM. At the indicated times, the medium was \\\I, me~oheme-'~~I-hemopexin (righthatch, / / /) or [6sFe]meso- removed, the cells washed with ice-cold buffer, and scraped from the heme-hemopexin (left hatch, \ \ \) at 37 "C. After 15 min of incuba- dish. After homogenization and low speed centrifugation, the supertion, the cell-associated radiolabel was measured as described under natant was mixed with Percoll, centrifuged, and fractions were ana"Materials and Methods." lyzed as described under "Materials and Methods." The profiles shown are those obtained after 5 (open circles) and 30 (closed circles) cells (Fig. 6), affinity per se should not affect determination min of incubation with [66FeJrnesoheme-hemopexin, and after 5 (open of the maximum number of binding sites. Thus, it appears squares) and 30 (closed squares) min of incubation with 56FeTPPSthat additional differences exist in theway the two complexes hemopexin. Fraction 1 is the lightest and fraction 25 the heaviest are handled by the cells and in the events subsequentto their fraction. [HEME-HEMOPEXIN], nY

binding. Short-term incubations (e.g. 15 min) are generally considered to represent binding, but some endocytosis or recruitment of receptors to the surface may occur and contribute to the differences in thenumber of binding sites. Both endocytosis of heme-hemopexin and thepresence of an internal pool of hemopexin receptors have been noted.5 That events subsequent to binding of one complex to cells differ from those of the other are supported by the observation that the amount of cell-associated radiolabel from "FeTPPShemopexin at 37 "C increases linearly but for a shorter time (90 min) and at a rate about one-third that of["Fe]mesoheme-hemopexin (Fig. 6 B ) .As expected from the recycling of hemopexin to the medium (I), more iron-tetrapyrrole than protein is associated with the Hepa cells from both FeTPPShemopexin and mesoheme-hemopexin at all concentrations is tested. However, the ratio of "Fe-ligand to '251-hem~pe~in consistently much lower for FeTPPS-hemopexin than for mesoheme-hemopexin (Fig. 7). To further compare the cell transport of these two irontetrapyrroles, isopycnic gradient centrifugation analyseswere carried out in parallel experiments. As shown in Fig. 8, after 5 min at 37 "C radiolabel from boththe [55Fe]mesohemehemopexin complex and the "FeTPPS-hemopexin complex is found in less dense (density < 1.03 g/cc) fractions containing plasma membranes, the hemopexin receptor and MHBP (3, 10, 22). However, by 30 min radiolabel only from [55Fe] mesoheme is transferred efficiently to intracellular compartments of density -1.04 and 1.08 g/cc. The former region of such gradients has been reported (19) to include endosomes (which contain MHBP'). The latter region has been shown to contain mitochondria and heme derived from heme-hemopexin (22). The 55FeTPPS remains associated throughout this time with the lighter fractions containing plasma membranes, probably as either FeTPPS-hemopexin bound to the receptor or as "FeTPPS-MHBP since MHBP binds FeTPPS.' In view of the differences in subcellular distribution with time, not only is FeTPPS not efficiently transferred to MHBP but additional eventsnecessary for intracellular transport areimpeded. A. Smith, E. Cohen, and R. Hunt, submitted for publication.

Our working hypothesis is that heme-hemopexin complexes bind to the hemopexin receptor resulting in the transfer of heme released from hemopexin to MHBP followed by recycling of hemopexin. Thus, takentogether the results in Table 11and Figs. 6-8 suggest that while FeTPPS-hemopexin binds to thereceptor, there is impairment of subsequent events such as heme release or transport in part due to the chemical differences between FeTPPS and heme. Of special note is the additional evidence provided by this work that conformational changes in hemopexin induced by heme binding and reflected by CD and absorbance spectra are clearly important for its function. This is manifest, for example, in the higher affinity of heme-hemopexin than of FeTPPS- or apo-hemopexin for the receptor. FeTPPS appears to produce some, but not all, of the conformational changes in hemopexin that heme does, enabling this more detailed examination. The picture that emerges from this work and other recent results4 indicate that at least three changes in conformation occur in hemopexin upon binding heme. For convenience, the changes currently identified have been designated as types I, 11, and 111. Type I is associated with alterations in the environment of tryptophan residues (probably in domain I) and is reflected by characteristic changes in both theabsorbance and CD spectra of hemopexin. For this type of change, a bis-histidyl coordination complex is a necessary but notsufficient condition. Since the intrinsic positive ellipticity at 231 nm of hemopexin resides in domain I1 (7), and since no change in either absorbance or CD occurs in domain I upon binding heme in the presence or absence of domain 11 (7), the shiftsin absorbance and CD spectra require an intacthinge region linking the two domains of hemopexin and result from the interaction between the two domains which alters upon heme binding. The Type I1 conformational change confers resistance to proteolysis of the hinge region. This Type I1 conformational change, seen upon binding FeTPPS, requires only the coordination of two histidine residues with the heme-iron but does not produce the characteristicalterationsin the absorbance or CD spectra of hemopexin. The Type I11 conformational change increases the affinity of hemopexin for its receptor. Type 111 requires a

Determinants of Hemopexin Transport Functions bis-histidyl coordination of the heme and isassociated with a compaction of domain 1. However, as shown elsewhere: full binding to the receptor requires both domains and bound heme but not an intact hinge region. The increase in the compactness of domain I caused by heme, but not FeTPPS, appears to be particularly important. Moreover, although FeTPPS-hemopexin interacts with the hemopexin receptor, the subsequent events in hemopexin-mediated heme transport, including the transfer of the heme from hemopexin to MHBP mediated by the hemopexin receptor and the intracellular transport of heme are perturbed. Studies to further characterize these events using this and other heme analogs are in progress. Acknowledgments-We are grateful to Victor Deaciuc for his excellent technical assistance, to Dr. Gerard Jansen for aiding us in carrying outpreliminary sucrose gradient analyses, and toDr. Preston Hensley (GeorgetownUniversity Medical Center, Washington,D. C.) for providing the program for analyzing untransformed binding data and for helpful discussions on its application.

REFERENCES 1. Smith, A., and Morgan, W. T.(1979) Biochem J. 182,47-54 2. Smith, A., and Morgan, W. T. (1981) J. Biol. Chem. 2 5 6 , 1090210909 3. Smith, A., and Morgan, W. T. (1984) J. Biol. Chem. 259,1204912053 4. Morgan, W. T. (1976) Ann. Clin. Res. 8,Suppl. 17, 223-232 5. Morgan, W. T., Sutor, R. P., and Muller-Eberhard, U. (1976) ActaBiophys. Biochim. 434,311-323

5229

6. Morgan, W. T., and Vickery, L. E. (1978) J. Biol. Chem. 2 5 3 , 2940-2945 7. Morgan, W. T., andsmith, A. (1984) J.Biol. Chem. 2 5 9 , 1200112006 8. Smith, A., and Morgan, W. T. (1984) Protides Biol. Fluids Proc. Colloq. 3 1, 219-224 9. Takahashi, N., Takahashi, Y . , and Putnam, F. W. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 73-77 10. Smith, A., and Morgan, W. T. (1985) J. Biol. Chem. 2 6 0 , 83258339 11. Gibbs, E., Skowronek, W. R., Morgan, W. T., Muller-Eberhard, U., and Pasternack, R.F. (1980) J.Am. Chem. SOC. 102,39393944 12. Hrkal, Z., Kalousek, T., and Vodrazka, Z. (1980) Znt. J. Biochem. 12,619-624 13. Hrkal, Z., and Muller-Eberhard, U. (1971) Biochemistry 1 0 , 1746-1750 14. Adler, A. D., Longo, F. R., Kampas, F., and Kim, J. (1970) Znorg. Nucl. Chem. 32,2443-2445 15. Seery, V. L., Hathaway, G., and Muller-Eberhard, U. (1972) Arch. Biochem. Biophys. 160,269-272 16. Brown, S . B., and Lantzke, I. R. (1969) Biochem. J. 1 1 5 , 279285 17. Miles, E. W. (1977) Methods Entymol. 47,431-442 18. Morgan, W. T., and Muller-Eberhard, U. (1976) Arch. Biochem. Biophys. 1 7 6 , 431-441 19. Dickson, R., Hanover, J. A., Willingham, M. C., and Pastan, I. (1983) Biochemistry 2 2 , 5667-5674 20. Kaminsky, L. S., Byrne, M. J., and Davison, A. J. (1972) Arch. Biochem. Biophys. 150,355-361 21. Morgan, W. T., and Muller-Eberhard, U. (1972) J. Biol. Chem. 247,7181-7187 22. Smith, A. (1985) Biochem. J.231,663-669