EXPERIMENTAL AND MOLECULAR PATHOLOGY 43, 64-73 (1985)

Doxorubicin and Covalently Crosslinked Doxorubicin Derivatives Binding to Purified Cardiac Thin-Filament

Proteins in Vitro

WILLIAM LEWIS,~,~ KEVIN BECKENSTEIN, LAWRENCE SHAPIRO, AND

SAUL PUSZKIN’

‘Laboratories of Molecular Pathology, Depurtment of Pathology. Mount Sinai School of Medicine of the City University of New York, Fifth Avenue und 100th Street, Nen, York, Nen, York 10029

Received September 25, 1984, and in revised form Februury 4, 1985

The binding of cardiac actin and tropomyosin to a cardiotoxic antineoplastic agent, doxo- rubicin, and its covalently crosslinked derivatives was investigated. The primary amino group of the daunosamine moiety of doxorubicin was blocked with fluorescein isothiocya- nate. This doxorubicin derivative did not bind to Sepharose which was conjugated with cardiac actin. A doxorubicin dimer was made by covalently crosslinking one doxorubicin molecule to another identical doxorubicin molecule through the free amino group of each daunosamine moiety. This derivative demonstrated mobility different from parent doxoru- bicin on thin-layer chromatography, different elution pattern by column chromatography, and did not show binding affinity for actin. Exploring other purified thin-filament proteins, it was found that doxorubicin did bind to tropomyosin when gel filtration was performed on the protein drug mixture. The ability of tropomyosin to form paracrystal in vitro was not disturbed by a variety of concentrations of doxorubicin. These data support the concept that the doxorubicin solitary free amino group is the site which is responsible for this ligand to bind to actin and may relate to its cardiotoxic effects. ‘e 1985 Academic Press. Inc.

INTRODUCTION

Doxorubicin (Adriamycin, ADR), a widely used anthracycline antineoplastic drug, is limited by a dose-related, severe, irreversible cardiotoxicity (Lefrak et al., 1973). This iatrogenic entity is manifest by a variety of ultrastructural changes observed clinically in humans and experimentally in animals (Benjamin et al.,

1978; Jaenke, 1974, 1976) in which heart cell organelles, including mitrochondria, thin filaments, and Z-bands, are involved (Duarte-Karim et al., 1976; Jaenke, 1976). Although this compound is used clinically and has been investigated phar- macologically (Bachur et al., 1976; Tong ef al., 1979; Trouet et al., 1979) the manner by which biochemical derangements in cardiac contractile apparatus occur has been investigated using purified myocardial contractile proteins (Lewis et al., 1982) and cultured myocardial cells (Lewis et al., 1983). To gain further insight into the molecular alterations produced by this drug, the ability of ADR to bind to proteins of the cardiac thin filament was tested and ADR was altered chemically to inhibit this binding.

Adriamycin is composed of two moieties: Adriamycinone, an anthracycline ring structure, and daunosamine, an amino sugar. The drug is brilliant red in its lyophilized state and in solution of neutral pH range, bright orange below pH 5, and purple above pH 8, implying 2 pK,. The present study implicates the binding site of ADR for macromolecules to be the free amino group of daunosamine. This amino functional group is necessary for interaction with cardiac actin (Fig. la).

? To whom all correspondence should be addressed at: Department of Pathology. UCLA Medical Center, Los Angeles, Calif. 90024.

64

0014-4800/85 $3.00 CopyrIght 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.

DOXORUBICIN BINDING TO CARDIAC PROTEINS 65

MATERIALS AND METHODS

All reagents were analytical grade I. Adriamycin was donated by Adria Phar- maceutical, Columbus, Ohio. Heart muscle used to isolate and purify contractile proteins was obtained fresh from normal mongrel dogs. Sephadex and Sepharose products were from Pharmacia Fine Chemicals, Piscataway, New Jersey. [i4C]ADR (sp act = 25 mCi/mm) was obtained from Stanford Research Institute, Stanford, California.

Buffer A: 0.40 M KCl, 0.05 M MES3 14 mM 2 ME, 0.4 mM ATP, pH 6.8. Buffer B: 2 mM MES, 0.2 mM ATP, 14 mM 2 ME, 0.2 mM CaCl?, pH 7.2.

Preparation of dog heart muscle actin. The techniques used were modified from those described previously (Spudich and Watt, 1971). Fresh dog heart (150 g wet wt) trimmed, minced, and washed in saline, was blended for 10 set in a Waring blender in 200 ml of Buffer A. The frothy suspension was stirred for 10 min, centrifuged at 10,OOOg for 5 min, and residue was washed consecutively with three vol of 0.05 M bicarbonate buffer, pH 8, for 15 min followed by a wash with 20 vol of 1 mM EDTA, pH 7.0, and with 10 vol of deionized, distilled water to remove EDTA. The residue was extracted with 2 vol of ice-cold acetone for 5 min, air-dried overnight and extracted at 4°C for 16 hr with 300 ml of Buffer B. The extract was filtered and the supernatant solution adjusted to pH 7.6. The suspension was clarified by ultracentrifugation at 100,OOOg for 20 min and solid KC1 and 1 M MgCl, were added to a final concentration of 100 and 1 mM, respectively. To remove tropomyosin, solid KC1 was added to a final concentra- tion 0.6 M and the solution allowed to stand at 4°C for 0.5 hr. The solution was centrifuged at 100,OOOg for 1.5 hr; the gelatinous clear pellet was suspended in and dialyzed against Buffer B. The tinal protein solution was clarified by ultra- centrifugation and stored frozen in 10% sucrose at -20°C until use.

Preparation of dog heart tropomyosin. The techniques used have been de- scribed previously (Bailey, 1948). The final protein solution was clarified by ul- tracentrifugation at 125,OOOg for 1 hr at 4°C and stored frozen at -20°C in 10% sucrose.

Coupling of G-actin to CNBr-Sepharose 4B. Before conjugation, 2 ml of 6 mg/ ml G-actin was dialyzed overnight in Buffer B, centrifuged at 100,OOOg for 20 min, and the supernatant added slowly to 5 ml of swollen CNBr-Sepharose 4B gel beads with constant, gentle agitation as described by Puszkin et al. (1983). The material was mixed in an orbital shaker for 2 hr at room temperature. To block all unreacted groups, the beads were washed, neutralized, and an equal volume of 0.5 M ethanolamine was added. The Sepharose beads were stored at 4°C in 0.05 M bicarbonate buffer, pH 8. Approximately 3 ml of the gel beads was applied to a small column to yield a gel pack of 1 x 2.5 cm.

Conjugation of ADR tofluorescein isothiocyanate. Fluorescein isothiocyanate (0.1 mmole) was placed in 5 ml of Buffer A. To this material 0.1 mmole of ADR was added at 25°C and stirred in darkness for 4 hr. This mixture was centrifuged at 17,OOOg for 10 min and 200 ~1 of supernatant solution was applied to the actin affinity column and eluted with Buffer B. Fractions of 1 ml were collected and each fraction was monitored spectrofluorimetrically.

Preparation of ADR derivative using dimethylsuberimidate as covalent cross-

3 Abbreviations: MES = morpholinoethanesulfonic acid; 2 ME = mercaptoethanol.

66 LEWIS ET AL.

linker. ADR was mixed with dimethylsuberimidate (DMS; Pierce, Rockford, Ill.) (Fig. la,b) in solution containing 75 mM triethanolamone (Fisher Scientific, Pitts- burgh, Pa.) in deionized, distilled water. DMS (50 pmole) was added slowly until it dissolved. Adriamycin (50 pmole) was added and stirred at room temperature in darkness for 1 hr. An additional 50 p.mole ADR was added and the mixture stirred for 2 hr at room temperature in darkness.

Thin-layer chromatography of ADR and derivatives. Adriamycin standard so- lutions were made (1 mg/ml ADR) in double-distilled water; 1.5 ~1 was used as a standard. Equivalent volumes of DMS-ADR derivative were plated on Eastman TLC plates (Eastman Chemical, Rochester, N.Y.), and allowed to air-dry in dark- ness for 6-8 hr. Chromatography was performed using a solvent consisting of l- propanol:ethyl acetate:water (7:1:2, v/v) (Solvent A). Thin-layer plates were re- moved from chromatography tanks, air-dried in a fume hood, and examined vi- sually in daylight and under a fluorescent lamp.

Silica gel column chromatography. A 20 x 1.5-cm chromatography column (Pharmacia) was packed by applying 20 g BioSil A (Bio-Rad, Richmond, Calif.) in 100 ml of Solvent A described above. A 0.5-ml aliquot of the DMS-ADR derivative was placed on the column and eluted with Solvent A. Two-milliliter fractions were collected by an automated fraction collector (Buchler Instruments, Fort Lee, N.J.), and absorbance monitored at 500 nm in a spectrophotometer.

Gel ajjjnity chromatography of ADR derivatives on actin-coupled CNBr-Sepha- rose 4B. The unpurified derivative mixture from the DMS-ADR reaction was added (200 ~1) to a 3 x l-cm actin-conjugated CNBr-Sepharose 4B column to remove unreacted ADR. Fractions of 1 ml were collected and absorbance of each fraction was monitored at 500 nm in a Beckman DB spectrophotometer. After 8 ml of elution, buffer was changed to Buffer A. After this, the gel was equilibrated again with Buffer B and 0.5 ml ADR (1 pmole) was applied to the column. Frac- tions of 1 ml were collected and after 8 ml, the buffer was changed to Buffer A. Absorbance of the fractions was monitored photometrically at 500 nm.

Binding of ADR to purified cardiac tropomyosin. One milligram per milliliter of purified cardiac tropomyosin which was frozen in 10% sucrose was then thawed, dialyzed against Buffer B, and clarified by ultracentrifugation at 40,OOOg for 1 hr. An aliquot (1 ml) of the protein was then incubated with 1 pmole ADR containing 0.01 tKi of [14C]ADR for 1 hr at room temperature in darkness. A 20 x l-cm column containing G-25 Sephadex (Pharmacia) was built and equilibrated with Buffer B. The incubation mixture was passed through the column, eluted in l-ml fractions using Buffer B and changed to Buffer A after 8 ml of elution.

H

t NH,

b

FIG. 1. Structure of doxorubicin (Adriamycin) mol wt = 543 (a). Arrow indicates site for cross- linking with dimethylsuberimidate (b).

DOXORUBICIN BINDING TO CARDIAC PROTEINS 67

Aliquots of each fraction (100 pJ) were counted for radioactivity using a Tricarb scintillation spectrometer and Aquasol II scintillation cocktail. Absorbance at 280 nm was monitored on the balance of each fraction.

Negative-staining electron microscopy of tropomyosin paracrystals. Tropo- myosin aliquots were dialyzed for 16 hr at room temperature against 50 mM MgClz Tris-HCl, pH 8.0, with and without ADR, and were examined by negative- staining electron microscopy. Ten microliters of 1 mg/ml tropomyosin was added to 90 pJ of Buffer B containing from lop6 to 10m4 M ADR. These mixtures were incubated individually at room temperature for 1 hr and aliquots were separately removed for examination by electron microscopy. Samples were deposited on Formvar copper-coated grids and allowed to stand for 20 sec. The excess solution was removed and negative staining was performed using 1% uranyl acetate. Grids were dried and observed with a JEOL 100 B electron microscope at 80 kV.

Sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis. Polypeptides were resolved in 7-15% gradient SDS-polyacrylamide slab gels. Gels were al- lowed to run for 16 hr at room temperature with a constant amperage of 7.5 mA per gel. Gels were fixed, stained with Coomassie brilliant blue R 250, and de- stained by diffusion in diluted fixative solutions (Laemmli, 1970).

RESULTS

Purification of Thin-Filament Proteins

Canine cardiac actin and tropomyosin were judged 90-95% pure by slab gel electrophoresis (Fig. 2) and biophysically and biochemically showed character- istics of the respective proteins as reported previously (Bailey, 1948; Go11 et al., 1972; Katz et al., 1966; Robson et al., 1970; Spudich and Watt, 1971).

The ADR-fluorescein compound was passed through the actin-CNBr-Sepha- rose 4B column as described. The results shown (Fig. 3) indicate that the fluo- rescein isothiocyanate-coupled ADR did not bind to the actin-CNBr-Sepharose 4B affinity column and eluted with the running buffer.

An ADR derivative was formed by using DMS, a substance which crosslinks primary amino groups. When DMS-ADR was applied to the actin affinity

1 2 3

FIG. 2. SDS-Polyacrylamide gel electrophoresis of cardiac muscle proteins. Slab gels were created with a continuous-gradient (7- 15%) polyacrylamide. Proteins were applied, electrophoresis was per- formed, and gels were fixed, stained with Coomassie blue. and gels were destained by diffusion. Lane 1: heart muscle tropomyosin (c) 2.5 ug; lane 2: heart muscle a-actinin (a) 30 pg; lane 3: heart muscle actin (b) 50 pg.

68 LEWIS ET AL.

FIG. 3. Affmity chromatography with actin-CNBr-Sepharose 4B and a fluorescein isothiocyanate ADR derivative. ADR was conjugated to fluorescein isothiocyanate. This ADR derivative (0.5 ml of lo-” M) was passed through an actin-coupled CNBr-Sepharose 4B afftnity column and collected fractions were monitored for fluorescence. A fluorescent peak was detected in the earlier eluted fractions. Little fluorescence could be detected subsequently (buffer change to Buffer A indicated by downward arrowhead; 0 indicates fluorescein fluorescence; 0 indicates ADR fluorescence).

column, a discrete red coloration was noted in fractions 5-8 eluting with the running buffer. When the elution buffer was changed to Buffer A, another faint pink coloration of the eluant was noted to correspond to a second well-defined absorbance peak. This latter elution pattern was similar to that of control ADR (Fig. 4). Aliquots of the DMS-ADR were analyzed by thin-layer chromatography, and found to migrate with an Rf of 0.6-0.65; aliquots of the trailing (ADR) frac- tions had Rfof OS-O.55 (Fig. 5). This latter Rfwas similar to migration of control ADR on thin-layer chromatography.

To separate DMS-ADR derivative from ADR, a column chromatography system was developed employing a silica gel matrix which was similar in chemical composition to material used in the thin-layer plate. A sharp absorbance peak eluted in 35 ml. After 60 ml of elution, a second peak appeared between 60 and 130 ml of elution buffer. As a control, native ADR (0.4 mg) was chromatographed and yielded a solitary broad peak which eluted after 70 ml of solvent (Fig. 6).

d 0.2

c

8 14 Ln ; 2 0.1 s f

a

‘i’a, i

1 I

0 4 6 12 16 20 volume (ml.)

FIG. 4. Afftnity chromatography of ADR-DMS derivative on actin coupled to CNBr-Sepharose 4B. One-milliliter aliquot of the DMS-ADR was applied to the actin affinity column (3 x 1 cm column CNBr-Sepharose actin gel). Fractions were eluted in Buffer B and absorbance of each fraction was monitored photometrically at 500 nm. After 8 ml of elution, the buffer was changed to Buffer A (indicated by arrow) and collection of fractions continued as before. The first peak eluted in with the running buffer; the second peak eluted with Buffer A containing K+. The column was washed with 50 ml of Buffer A followed by equilibration with Buffer B. ADR (0.5 ml, 1 pmole) was applied and eluted in Buffer B, as above. After 8 ml, buffer was changed to Buffer A and elution continued. Absorbance of each fraction was monitored photometrically at 500 nm. (0 ADR derivative: 0 na- tive ADR).

DOXORUBICIN BINDING TO CARDIAC PROTEINS 69

1 2 3 4 5 6

14 12 10 9 7 5

fraction number

FIG. 5. Thin-layer chromatography of ADR and derivatives. ADR and its derivative were prepared in aqueous buffers. Aliquots of ADR were applied to thin-layer plates, as were equivalent volumes of DMS-ADR derivative. Chromatography proceeded in darkness for 8 hr using a propanol:ethyl- acetate:water (7:1:2 v/v) solvent system. Spots 1, 2. and 3 are aliquots of fractions 14, 12, and 10. respectively, of elution profile shown in Fig. 3. Spots 4, 5. and 6 are fractions 9, 7. and 5, respectively, from elution profile of DMS-ADR (0 in Fig. 3).

Binding of ADR to Cardiac Tropomyosin

When a mixture of heart muscle tropomyosin was incubated with a threefold molar excess of ADR containing [14C]ADR and passed through a G-25 Sephadex column, two radioactive peaks of [t4C]ADR eluted. The first radioactive peak eluted in the void volume. A second broad peak eluted after the buffer change. Absorbance at 280 nm was obtained on the fractions and revealed a protein peak which comigrated with the first peak of radioactivity; no other 280-nm absorbance peak was detected after the buffer change (Fig. 7). Cardiac tropomyosin dem- onstrated characteristic paracrystal formation with periodicity of approximatel; 380 A which was undisturbed by the presence of ADR (Fig. 8).

DISCUSSION

The nature of ADR binding to cardiac actin and tropomyosin may shed light on biochemical phenomena involving ADR and these sarcomeric constituents and

GO.2

:

2 ,o. 1

;T ; ; ;,

s

/ i

E g

lzz . . . . . . .,, ,. i ‘....

$0 40 120 200 volume effluent--ml.

FIG. 6. Silica gel column chromatography of derivatized DMS-ADR. A Biosil A 20 x 6-cm column was packed and 2 ml of DMS-ADR was applied and eluted with Solvent A. Fractions (2 ml) were obtained for 200-ml elutions volume and monitored at 500 nm photometrically. Two discrete absor- bance peaks were visible: each one was a sharp peak eluting in the first 32 ml and a broad, diffuse peak eluting from 60- 130 ml. Dotted line indicates elution pattern of 0.4 mg of ADR. Only one broad peak is present in the ADR elution pattern.

70 LEWIS ET AL.

volume--ml.

FIG. 7. Binding of [t4C]ADR to tropomyosin using gel filtration chromatography. A 1 x 20-cm column of G-25 Sephadex was packed (equilibrated in Buffer B). One milligram of tropomyosin is incubated with 0.5 ml of 10m6M ADR containing 0.01 uCi [14C]ADR. The ADR-tropomyosin mixture was eluted in Buffer B and fractionated. Aliquots (100 (~1) of each fraction were counted for radio- activity in a scintillation spectrometer. The balance of the material was monitored at 280 nm photo- metrically. Arrow indicates change to Buffer A. (0 absorbance at 280 nm; Cl cpm x lo-*).

.l Alurn

1 cm

.05 l.lm

1cm

FIG. 8. Negative-staining electron micrograph of cardiac tropomyosin paracrystals. (A) formed in the presence of 1 pmole ADR. (B) formed in absence of ADR. Ten microliters of 1 mg/ml tropomyosin were deposited on Formvar copper-coated grids and allowed to stand 20 sec. Excess solution was removed and negative staining performed in 1% uranyl acetate dried and observed with JEOL 1OOB electron microscope. Both ADR-treated (A) and -untreated (B) tropomyosin demonstrate similar paracrystal formation with 380 A register, characteristic of tropomyosin.

DOXORUBICIN BINDING TO CARDIAC PROTEINS 71

relate to cardiomyopathy. Disruption of the thin filaments may be effected by the interaction of cardiac contractile proteins with the drug. Previous results from our laboratories showed that interacting ADR with G-actin in low-ionic-strength buffers resulted in formation of unusual bundles of actin visualized by negative- staining electron microscopy; biochemically, we found Mg*+-actin-stimulated ATPase by myosin in the presence of ADR in the system (Lewis ef al., 1982).

Using a cultured cardiac myocyte system (CMC), cytoplasmic protein synthesis was previously examined. Effects of ADR on protein synthesis in this system indicated a decrease in CMC protein synthesis in cells exposed to ADR for 24 hr (Lewis et al., 1983). In correlative ultrastructural studies on ADR treated CMC, clear zones containing short microfilaments were noted in these cells.

Binding a compound to CNBr-coupled-Sepharose 4B requires a primary amino group in the ligand for successful conjugation. Free amino groups from lysine, arginine, and N-terminus of polypeptides would account for the binding of pro- teins to the gel matrix. Adriamycin has one free amino group on daunosamine which accounts for the binding of ADR to the CNBr-Sepharose 4B beads. We previously reported binding studies of F- and G-actin and cw-actinin with ADR- CNBr Sepharose 4B (Lewis et al., 1982). Investigating this point, ADR was con- jugated stoichiometrically with fluorescein isothiocynate to block ADR’s free amino group.

Dimethylsuberimidate crosslinked ADR. Minimal breakdown products were detected chromatographically despite the sensitivity of ADR to light. Precautions were taken during the experimental procedures to avoid ultraviolet exposure. Reactions were performed in darkness with gentle agitation. Separation of un- conjugated ADR from the dimerized product was based on small, significant differences in mobilities between these molecules demonstrated by thin-layer chromatography. This difference in partitioning between DMS-ADR and ADR was noted and the principle applied to a column chromatograph which used a column bed made of the same inert matrix as in the thin-layer system and the same solvents. In affinity chromatography studies, the drug with modified dau- nosamine moiety (either with fluorescein isothiocyanate or the DMS-ADR dimer) similarly eluted from the affinity chromatography column without binding to the proteins which were coupled to the matrix.

ADR binding to actin appears to be mediated by the interaction of the amino group present in the daunosamine moiety of ADR. When DMS-ADR was formed by crosslinking, the ADR primary amino functional groups became part of an amide linkage and thus became chemically unavailable for interaction with actin linked to the gel matrix. In reciprocal ways, binding ADR to the gel matrix al- lowed proteins applied to the column to elute unretarded (Lewis et al., 1983). It is possible that ADR may interfere with contractility and the maintenance of integrity of the thin filament in vivo. If thin-filament protein homeostasis is dis- turbed by ADR (by aberrant actin monomer polymerization) and cannot be com- pensated by normal addition of new actin monomers to the dynamic thin filament (in a steady state), structural integrity of the sarcomere could be jeopardized.

We have shown previously how ADR alters some biophysical and biochemical properties of cardiac actin in vitro (Lewis ef al., 1982). Such effects may occur in ADR heart muscle disease and may account for sarcomeric disarray seen ul- trastructurally in ADR cardiotoxicity. If this ADR effect is mediated at the level of genomic expression, protein synthesis (particularly contractile protein syn-

72 LEWIS ET AL.

thesis) in the heart cell may be altered radically. Anthracycline-derivative com- pounds differing chemically from ADR have shown promise as clinically effective antineoplastics (Tong et al., 1979; Trouet et al., 1979) and may demonstrate less toxic side effects.

The data presented support the importance of the ADR free amino group in coupling to an organic matrix and binding macromolecules, particularly cardiac thin-filament proteins: actin and tropomyosin. Examination of in vitro tissue cul- ture in vivo systems using ADR and derivatives of ADR may help explain altered sarcomeric structure in ADR cardiotoxicity. Such studies may lead to develop- ment of new anthracycline derivatives which have higher therapeutic index and less cardiotoxicity.

ACKNOWLEDGMENTS William Lewis is recipient of a Clinical Investigator Award from the NHLBI, K8 HL 01295. This

work was supported in part by the California Institute for Cancer Research, Los Angeles, California and in part by the Cancer Research Coordinating Committee, Berkeley, California.

REFERENCES BACHUR, N. R., STEELE, M., MERIWETHER, W. D., and HILDEBRAND, R. D. (1976). Cellular phar-

macodynamics of several anthracycline antibiotics. J. Med. Chem. 19, 651-654. BAILEY, K. (1948). Tropomyosin: A new asymmetric protein component of the muscle fibril. Biochem.

J. 43, 271-279. BENJAMIN, R. S., MASON, J. W., BILLINGHAM, M. E., and DANIELS, J. R. (1978). Doxorubicin car-

diomyopathy: Evaluation by phonocardiography. endomyocardial biopsy and cardiac catheteriza- tion. Ann. Intern. Med. 88, 168-175.

CHAIRES, J. B., DATTAGU~A. N., and CROTHERS, D. M. (1982). Self-assocation of daunomycin. Bio- chemistry 21, 3927-3932.

DUARTE-KARIM, M., RUYSSCHAERT, J. M., and HILDEBRAND, J. (1976). Affinity of adriamycin to phos- pholipids: A possible explanation for cardiac mitochondrial lesions. Biochem. Biophys. Res. Commun. 71, 658-663.

GOLL. D. E., SUZUKI. A., TEMPLE, J., and HOLMES, G. R. (1972). Studies on a-actinin: Localization at the Z-bands of muscle sarcomeres. J. Mol. Biol. 67, 469-488.

JAENKE, R. S. (1974). An anthracycline induced cardiomyopathy in rabbits. Lab. Invest. 30, 292- 304.

JAENKE, R. S. (1976). Delayed and progressive myocardial lesions after adriamycin administration in the rabbits. Cancer Res. 36, 2958-2966.

KATZ, A. M., REPKE, D. I., and COHEN, B. R. (1966). Control of the activity of highly purified cardiac actomyosin by Ca+ +, Na+ , K+. Circ. Res. 19, 1062-1070.

LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacterio- phage T4. Nature (London) 227, 680-685.

LEFRAK, E. A., PITHA, J., ROSENHEIM, S., and GOITLIEB, J. A. (1973). A clinico-pathologic analysis of adriamycin cardiotoxicity. Cancer 32, 302-314.

LEWIS, W., KLEINERMAN, J., and PUSZKLN. S. (1982). Doxorubicin interaction with cardiac contractile proteins. Circ. Res. 50, 547-553.

LEWIS, W., GALIZI, M., and PUSZKIN, S. (1983). Compartmentalization of adriamycin and daunomycin in cultured chick myocardial cells: Effects on synthesis of contractile and cytoplasmic protein synthesis. Circ. Res. 53, 352-362.

PEPE, F. A. (1971). In “Subunits in Biological Systems” (S. N. Timasheff and G. D. Fasman, eds.). Vol. 5, pp. 323-353. Dekker, New York.

PUSZKIN, S., PUSZKIN, E., MAIMON, J., ROUNAULT, C., SCHOOK, W., ORES, C., and ROSENFIELD,

R. E. (1977). Alpha-actinin and tropomyosin interactions with a hybrid complex of erythrocyte- actin and muscle myosin. J. Biol. Chem. 252, 5529-5537.

PUSZKIN, S., ANDRES, A., ORES, C., LISANTI, M. P., and SCHOOK. W. J. (1983). Immunocytochemical characterization of clathrin-associated proteins (CAPS). I. Neuronal distribution of CAPS, com- ponents of clathrin coated vesicles. Cell Tissue Res. 231, 495-505.

DOXORUBICIN BINDING TO CARDIAC PROTEINS 73

ROBSON, R. M., GOLL, D. E., ARAKAWA, N., and STROMER, M. H. (1970). Purification and properties of a-actinin from rabbit skeletal muscle. Biochim. Biophys. Acfu 200, 296-318.

SPUDICH, J. A., and WATT, S. (1971). The regulation of rabbit skeletal muscle contraction. 1. Bio- chemical studies of the interaction of the tropomyosin-troponin complex with actin and the pro- teolytic fragments of myosin. J. Biol. Chem. 246, 4866-4871.

TONG, G. L., Wu, H. Y., SMITH, T. H., and HENRY, D. W. (1979). Adriamycin analogues. 3. Synthesis of N-alkalated anthracycline with enhanced efficacy and reduced cardiotoxicity. J. Med. Chem. 22, 912-918.

TROUET. T. (1979). Development of selective antitumoral drug-carrier complexes: Present status and prospects. Acta C/in. Be/g. 34, l-5.