Identification of a Cardiac Carnitine Binding Protein*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 257, No. 18, Issue of September 25, pp , 1982 Printed in U.S.A. Identification of a Cardiac Carnitine Binding Protein* (Received for publication, February 22, 1982) Carol R. Cantrell and Peggy R. Borum From the Departments of Biochemistry and Pediatrics, Vanderbilt University School of Medicine, Nashville, Tennessee The existence of a cardiac carnitine binding protein was demonstrated using an in vitro binding assay. The binding activity was solubilized with Triton X-1 from the pellet following a 59, x g centrifugation of rat ventricular homogenates. Preincubation at temperatures above 4 C or treatment with pronase significantly reduced the binding activity, suggesting that the activity was that of a protein. Cell fractionation studies suggested that the cardiac carnitine binding protein was associated with the plasma membrane fraction and that its activity was distinct from carnitine palmitoyltransferase, carnitine acetyltransferase, and carnitine translocase. Optimal binding required 6 min of incubation at 25 C. The binding of carnitine to the cardiac carnitine binding protein was saturable, and a dissociation constant of.7 pm for DL-carnitine was measured. L-Carnitine competed with DL-[methyl-3H]carnitine for binding to the cardiac carnitine binding protein, while competition by D-carnitine was much less effective. Binding was significantly inhibited when N-ethylmaleimide, iodoacetic acid, or mercuric chloride was present. Once DL-[ 3 H]carnitine was bound to the cardiac carnitine binding protein, radioactivity could be dissociated by a variety of mild treatments including dialysis, overnight incubation at 4 C, and application to a gel filtration column. Long chain fatty acids are unable to penetrate the inner mitochondrial membrane and are transported across the mitochondrial membrane as carnitine esters. The fatty acid of acyl-coa is linked to carnitine by carnitine palmitoyltransferase (1). Once inside the mitochondria, acyl-coa is reformed and undergoes #f-oxidation. Oxidation of long chain fatty acids is the major source of energy in the well oxygenated heart (2). Therefore, normal cardiac energy metabolism depends on carnitine. However, carnitine is not synthesized in the heart (3). Carnitine must be transported into cardiac tissue from the blood. Measurements of plasma carnitine in our laboratory are consistent with transport of carnitine into cardiac tissue against a concentration gradient (4). Active transport of carnitine into the heart has been demonstrated in several systems. Molstad et al. (5) measured a saturable uptake of carnitine in cultured human heart cells. The data of Vary and Neely (6) indicated a saturable component of carnitine transport in isolated perfused rat hearts. Bahl et al (7) demonstrated transport of carnitine in isolated adult rat heart my- * This work was supported by Grants AM 783 and AM 1619 from the National Institutes of Health and a grant from the Muscular Dystrophy Association of America. The costs of publication of this article were defrayed 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. ocytes which occurred against a concentration gradient and exhibited a saturable component. Molstad et al. (5) suggested that protein carriers exist on the plasma membrane of the myocyte to mediate uptake of carnitine by the heart cell. L- Carnitine in the growth media of cultured myocytes increased the activity of the carnitine uptake system. The rate of uptake was reduced when cycloheximide was included in the growth media (8). Preincubation with diphtheria toxin resulted in a decreased rate of uptake (9). Such evidence suggests that protein synthesis is necessary for carnitine transport, and, therefore, a specific binding protein may serve as a carrier for carnitine across the heart cell membrane. The experiments reported here were designed to investigate the possible existence of a cardiac carnitine binding protein and to distinguish it from proteins which are known to bind carnitine as substrate. EXPERIMENTAL PROCEDURES Sample Preparation-Ventricular tissue from male, adult, Sprague-Dawley rats was manually homogenized in Tenbroeck glass tissue grinders in 3 volumes of.1 M potassium phosphate buffer, ph 7.4. This homogenization and all subsequent steps were carried out at 4 C unless otherwise stated. The homogenate was centrifuged at 59, x g for 3 min. The pellet was resuspended in 3 volumes of.1% Triton X-1 in.1 M potassium phosphate buffer, ph 7.4, and incubated at 37 C for 3 min. This suspension was centrifuged at 15, x g for 6 min. To inhibit protease activity, 7 mg of phenylmethylsulfonyl fluoride was dissolved in 1 ml of 95% ethanol and added dropwise to each 5 ml of 15, x g supernatant. This solution was brought to 25 C and incubated at 25 C for 15 min. Following the phenylmethylsulfonyl fluoride treatment, the solution was stirred for 2 h with SM-2 Bio-Beads (1) at a concentration of.3 g/ml of solution in order to remove most of the Triton X-1. The solution was filtered to remove the Bio-Beads and the filtrate was diluted 1-fold with.1 M potassium phosphate buffer, ph 7.4. This dilution allowed Triton to be diluted below its critical micelle concentration. Triton could then diffuse out through the PM-1 membrane when the solution was concentrated to its original volume in an Amicon-stirred cell. This solution was dialyzed against.1 M potassium phosphate buffer, ph 7.4, for 48 h with daily changes of the buffer. Table I illustrates a study using (phenyl- 3 H]Triton X-1 to monitor the effectiveness of this methodology for removing Triton X- 1. The Triton concentration was reduced 18,-fold by this technique. The specific binding activity increased 2-fold following removal of Triton X-1 as seen in the table. When Triton X-1 was added back to the sample at a concentration of.8%, the binding activity was decreased by 75% which also indicated that Triton X-1 inhibited the binding activity. The term Triton X-1 treatment will be used to include both the solubilization of the protein with Triton X-1 and the removal of the detergent. Protein prepared by the Triton X-1 treatment containing approximately 1 mg of protein/ml in.1 M potassium phosphate buffer, ph 7.4, was used in all experiments to study the in vitro binding activity unless otherwise stated. Measurement of Sample Binding Activity-A sample of 5-1 pg of protein prepared by the Triton X-1 treatment was incubated in a final volume of.5 ml at 25 C for 6 mmin with 5 nmol of DL- [methyl-3h]carnitine (approximately 3 x 16 dpm). To correct for nonspecific binding, 5 nmol of L-carnitine hydrochloride were included in a second assay tube in addition to the components men- 1599

2 16 Cardiac Carnitine Binding Protein TABLE I Removal of Triton X-1 from CCBP preparation Ten #Ci of [phenyl- 3 H]Triton X-1 (1.56 mci/mg) were included in the.1% Triton X-1 used to solubilize protein from the heart homogenate pellet. Aliquots of the preparation were counted following each step to remove the Triton X-1. Another portion of the preparation at Steps 1, 2, and 5 was assayed for CCBP binding activity. The details of each step listed below are given under "Experimental Procedures." nmol carni- Treatment Total ton [ X-1 3 HITi- Triton bound/mg tine protein cpm % 1. Addition of Triton to 15.2 x pellet 2. 15, x g centrifugation of Triton suspension Supernatant Pellet 6.4 x x Stirred with Bio-Beads 244, and filtered 4. Dilution and concen tration 5. Dialysis tioned above but approximately 2% of the DL-carnitine counts remained bound to the protein. If 5 nmol of DL-carnitine were included in the assay, only 1% of the counts remained bound to the protein. Following the incubation, the sample was transferred to an Amicon MPS-1 micropartition system with a YMB membrane and spun at 152 x g for 3-5 min. The filter was washed five times with.5 ml of.1 M potassium phosphate buffer, ph 7.4, and transferred to a polyethylene scintillation vial. Following five washes, the number of counts retained by the filter remained constant. Therefore, all loosely associated or trapped counts were assumed to have been washed away. One ml of.1 M potassium phosphate buffer, ph 7.4, was added to the vial and the vial contents were mixed. Nine ml of ACS (aqueous counting scintillant from Amersham) were added and the vials were counted for radioactivity in a liquid scintillation spectrometer. Binding activity was measured in triplicate and results are expressed as nanomoles of carnitine bound/a28 unit of protein or per mg of protein: SB = (T - E)/P X C, where SB = specific binding, T = total counts/min present on the filter, E = total counts/min present on the filter when a 1-fold excess of L-carnitine hydrochloride was included in the incubation, P = A2so units of protein or mg of protein/ incubation volume, and C = conversion factor = nanomoles/cpm. The conversion factor allowed conversion of the bound counts/min to bound nanomoles of carnitine based on the calibration of a radioactive toluene standard and the radiospecific activity of the DLcarnitine. Enzyme Assays-Carnitine palmitoyltransferase was measured spectrophotometrically as previously described (11). Carnitine acetyltransferase was measured using a radioisotope method. The reaction tube contained 5 1l of the protein fraction (.3-.2 mg of protein) and 5 nmol of L-carnitine hydrochloride. The reaction was started with 1 id of a mixture which consisted of 2 volumes of 3.2 M [1-' 4 C]acetyl coenzyme A (radiospecific activity of 5 mci/mmol) and 1 volume of.1 mm acetyl coenzyme A and 1 volume of 1. M potassium phosphate buffer, ph 7., containing 2 mg/ ml N-ethylmaleimide. After incubation at 37 C for 3 min, the sample was applied to a column (5 x 35 mm) of Dowex 2X-8-C1-. The column was eluted into a polyethylene scintillation vial and washed with two.5-ml aliquots of cold H 2. Nine ml of ACS were added, and the vials were counted in a liquid scintillation spectrometer. 5'-Nucleotidase was measured as a plasma membrane marker using the method previously described (12). Succinic acid dehydrogenase was measured as a marker for mitochondria. Enzyme activity was assessed in a medium containing 1 mm sodium succinate, 1 mm potassium phosphate buffer, ph 7.4, 1 mm KCN, and the protein fraction containing.1-.5 mg of protein. Following a 5-min incubation at 37 C, neotetrazolium chloride was added to a concentration of 3 mm. Following an additional 3-min incubation at 37 C, 2 volumes of "stop solution" were added. The stop solution consisted of 5.6 g of Triton X-1, 8 ml of 4% formaldehyde, and 16 ml of.1 M formate buffer, ph 3.5, diluted to 94 ml with water. Absorbance was read at 55 nm. Results were expressed as AA/mg of protein. Non-collagen protein was solubilized by the method of Lilienthal et al. (13). Non-collagen protein was measured by the microbiuret procedure (14). Cell Fractionation Experiments--Fresh rat ventricular tissue was used for cell fractionation which was carried out according to procedure I of Jones et al. (15). This procedure involved a gentle homogenization and several centrifugations to remove contractile proteins. A more vigorous homogenization and series of centrifugations resulted in fractionation of the tissue into predominantly mitochondrial and plasma membrane fractions. The subcellular fractions obtained were assayed for the enzyme markers and the pellets were then subjected to the Triton X-1 treatment before measurement of binding activity. Materials-DL-[methyl- 3 H]Carnitine (1 Ci/mmol) was purchased from Amersham. The radiolabeled carnitine was 97-98% pure as determined by Amersham using thin layer chromatography on cellulose in three different solvent systems. [phenyl-3h]triton X-1 was purchased from New England Nuclear. Pronase, deoxyribonuclease, and ribonuclease were purchased from Sigma. Triton X-1 was purchased from Research Products International, Mt. Prospect, IL. The SM-2 Bio-Beads were purchased from Bio-Rad and phenylmethylsulfonyl fluoride was from Sigma. L-Carnitine hydrochloride and D-carnitine hydrochloride were generous gifts from Otsuka Pharmaceuticals of Japan. RESULTS The existence of CCBPI was first suggested from the results of an in vivo experiment. Rats were injected with 1 Ci of DL-[ 3 H]carnitine and killed 24 h later. Homogenates of the ventricular tissue were applied to Sephadex G-25. A small number of counts (less than 1% of the injected dose) was bound to the high molecular weight component which eluted from the column. An in vitro assay was then designed to detect a protein in cardiac tissue which bound carnitine. Rat ventricular tissue was homogenized and subjected to low speed centrifugation. Nonsignificant levels of specific binding activity were found in the soluble fraction. The pellet resulting from the 59, x g centrifugation of the ventricular homogenate specifically bound carnitine when the pellet was resuspended in.1 M potassium phosphate buffer, ph 7.4. The pellet was treated with various agents to solubilize the binding component as listed in Table II. Very little binding activity could be detected following treatment of the pellet with Lubrol at several concentrations or with 1. M KC1. These data suggest that the binding component was not simply trapped in the pellet or only loosely associated with membranes. Several times more binding activity was released from the pellet by.1% Triton X-1 than with the other treatments. A 2- to 3-fold increase of CCBP binding activity was measured in the pellet following solubilization by Triton X-1. Therefore, all subsequent experiments were conducted with preparations of Triton-solubilized protein from which the Triton X-1 had been removed as discussed under "Experimental Procedures." The technical details of the binding assay are given under "Experimental Procedures." Fig. 1A shows the counts/min which eluted with the high molecular weight component when a 15, x g supernatant was incubated with DL-[ 3 H]carnitine and then applied to Sephadex G-25. When a 1-fold excess of unlabeled L-carnitine was included in the incubation, the profile seen in Fig. 1B was obtained. Many fewer counts were associated with the high molecular weight component when excess nonradiola. beled carnitine was included in the incubation mixture which indicated that the binding activity was specific for carnitine. As the concentration of protein was increased in the binding assay, the specific binding activity increased in a linear relationship. The binding activity was also dependent on the 1 The abbreviation used is: CCBP, cardiac carnitine binding protein.

3 Cardiac Carnitine Binding Protein 161 TABLE II Release of CCBP binding activity from homogenate pellets Hearts were homogenized in.1 M potassium phosphate buffer, ph 7.4, and centrifuged at 15 x g for 3 min. The pellet was resuspended in 3 volumes of 1. M KCI or Lubrol at each concentration or.1% Triton X-1 and incubated at 37 C for 3 min. These samples were spun again at 15 x g for 3 min. The supernatants were dialyzed for 48 h against.1 M potassium phosphate buffer, ph 7.4, before assaying for CCBP binding activity. Treatment pmol/aao bound i.7 b x a L) KCI, 1. M 1.22 Lubrol.1%.4.5%.18 1.%.38 Triton X-1,.1% 4.56 FRACTION VOLUME (ml) FIG. 1. In vitro binding of DL-[ 3 H]carnitine to CCBP. Incubations were carried out as described under "Experimental Procedures" except that the incubation temperature was 37 'C. Following the incubation, the samples were applied to Pharmacia columns (.9 x 6 cm) of Sephadex G-25 in.1 M potassium phosphate buffer, ph 7.4. The columns were eluted with.1 M potassium phosphate buffer, ph 7.4. Aliquots of 1. ml were collected and monitored for absorbance at 28 nm and for radioactivity. The solid line represents absorbance at 28 nm, and the dashed line represents counts/min. A is the incubation mixture. B is the incubation mixture plus a 1-fold excess of unlabeled L-carnitine. temperature of incubation. Much less binding activity was measurable at incubation temperatures of 4 or 37 'C than at 2 or 25 C, and the optimal temperature for binding was 25 C. As the time of incubation was increased, the binding activity increased linearly up to 6 min of incubation. Beyond 6 min of incubation, the binding activity began to plateau. Fig. 2 shows that the binding activity increased linearly with increasing concentration of carnitine up to.5 pm and then began to plateau. Thus, the binding protein was saturable with respect to carnitine. An analysis of the data indicated a dissociation constant of.7 pm for DL-carnitine. The lack of availability of radioactive L-carnitine necessitated use of DLcarnitine. Insets a and b of Fig. 2 illustrate the Lineweaver- Burk plot and the Hanes plot of the data. The KM for the transport of carnitine into heart cells is 4.8 pm for cultured heart cells (5) and 6 pm for isolated heart cells (7). The protein nature of the CCBP was confirmed by the following experiments. If the preparation was boiled, no binding activity could be detected. If CCBP was incubated for 1 min at various temperatures prior to measurement of the binding activity, there was a progressive decrease in the binding activity measured as the preincubation temperature increased. The most dramatic decrease in binding activity occurred at preincubation temperatures above 35 C. There was no binding activity present when CCBP was preincubated at temperatures greater than 45 C. These results demonstrate that the binding activity of CCBP was heat-labile and extend the earlier results concerning the temperature dependence of the binding activity. When preincubated at temperatures greater than 25 C, a certain amount of the protein was altered such that less binding activity was measurable. CCBP was also preincubated for 3 min at 25 C with pronase (1 rng/ 1 mg of CCBP). Binding activity was measured following this preincubation, and the binding activity decreased 5% with pronase treatment. These data further suggest that CCBP is a protein. Preincubation with ribonuclease only slightly decreased the binding activity while preincubation with deoxyribonuclease slightly enhanced the binding activity. These results suggest that CCBP is not a polynucleotide. In order to determine the cellular location of the cardiac carnitine binding protein, fresh rat ventricular tissue was homogenized and fractionated by ultracentrifugation. Each of the resulting fractions was assayed for 5'-nuceotidase activity as a plasma membrane marker and succinic acid dehydrogenase as a marker for mitochondria. Carnitine palmitoyltransferase, carnitine acetyltransferase, and cardiac carnitine binding protein binding activity were also measured. The results of these measurements are seen in Table III. The three fractions included in the table were the three pellets obtained from the fractionation procedure and were the fractions which ZE Za WG Zo Z CARNITINE CONCENTRATION (M) FIG. 2. Dependence of CCBP binding activity on the concentration of carnitine. Aliquots of the cardiac carnitine binding protein preparation containing 5-1 ftg of protein were incubated with DL-[3Hcarnitine. The carnitine concentration varied from.125 to 4 /M in a preparation of 75% L-carnitine and 25% D-carnitine. A second set of triplicate tubes contained a 1-fold excess of L-carnitine hydrochloride for each concentration of DL-carnitine. Specific binding activity was calculated and the specific nanomoles of DL-[ 3 H]carnitine bound were plotted as a function of concentration of al-f 3 H]carnitine present. Inset a is the Lineweaver-Burk analysis of this data, and inset b is the Hanes plot of the data. S, concentration (em) of carnitine; B, binding activity in nanomoles/mg of protein/h.

4 162 Cardiac Carnitine Binding Protein TABLE III Subcellular distribution of CCBP Portions of each pellet fraction were-resuspended in.25 M sucrose, 5. mm histidine and this suspension was assayed for enzyme activities. Another portion of each pellet was treated with the Triton X- 1 treatment described under "Experimental Procedures," and the binding activity of CCBP was measured. Percentages given represent the amount of activity based on the total present in the three fractions included. Protein Pellet I Pellet H Pellet m Cardiac carnitine binding protein (nmol/mg protein) 5'-Nucleotidase (umol P 4 /mg protein) Succinic acid dehydrogenase (A/min/mg protein) Carnitine palmitoyltransferase (AA/min/mg protein) Carnitine acetyltransferase (jmuol carnitine/ao) (33%).16 (33%).316 (82%).263 (1%).371 (61%).3 (65%).616 (18%).58 (6%).3 (2%).18.1 (93%) (3.5%) (3.5%) contained significant amounts of these activities. These three fractions contained 85% of the specific binding activity of CCBP and 95% of the total protein. The mitochondria were predominantly in Pellet I and the plasma membranes were predominantly in Pellet II based on the marker enzymes assayed. Carnitine palmitoyltransferase and carnitine acetyltransferase are mitochondrial enzymes and were almost exclusively in the mitochondrial fraction or Pellet I. The activity of the cardiac carnitine binding protein was predominantly in the plasma membrane fraction. Pellet I contains 33% of the 5'-nucleotidase activity, indicating that plasma membranes are contaminating the mitochondrial pellet. It is likely that this plasma membrane contamination accounts for the CCBP binding activity also present in Pellet I. Carnitine translocase catalyzes a reversible exchange diffusion of acylcarnitine for carnitine across the inner mitochondrial membrane (16, 17). The previous experiment suggested that CCBP was associated with the plasma membrane fraction which implied that CCBP is not carnitine translocase. Furthermore, Pande and Parvin (18) have found that choline and betaine do not affect the activity of the carnitine translocase. However, as seen in the study described in the next paragraph, choline and betaine significantly compete for the binding of carnitine to CCBP. Therefore it is likely that CCBP is not carnitine translocase. The structural specificity of the binding activity was determined using structural analogs of carnitine in a competitive binding assay. Table IV illustrates that several of the analogs had a dramatic effect while others were much less effective in competing with DL-[ 3 H]carnitine for binding to CCBP. The most effective competitors were L-carnitine, betaine, choline, and N,N-dimethylglycine. y-butyrobetaine was included in earlier experiments at a concentration of 1. mm. It competed for 66% of the binding activity at that concentration. D-Carnitine did not significantly compete for the binding as seen in Table IV. The difference in the binding affinity for L-carnitine and D-carnitine is further substantiated in Fig. 3. D-Carnitine had very little effect on the binding of DL-[ 3 H]carnitine to CCBP when present at concentrations from 5 to 4 M. L- Carnitine maximally competed for 8% of the binding of the radiolabeled compound at a concentration of 1 fpm. Application of CCBP to Sephadex G-1, Sephacryl S-2, and Sepharose 2B resulted in each case in elution in the void volume of a peak which possessed CCBP binding activity. A small second peak which exhibited CCBP binding activity eluted from the Sepharose 2B column. This second peak had a molecular weight greater than that of thyroglobulin (Mr = 669,). These data suggest that CCBP is a very large complex. Since two peaks with binding activity eluted from Sepharose 2B, different forms of CCBP may be present. It is likely that the binding protein aggregates since centrifugation of a preparation at 15, x g for 3 h following removal of Triton resulted in a supernatant with only 25% of the binding activity present before centrifugation. Studies were carried out to look at the dissociative behavior of CCBP once carnitine is bound. CCBP and DL-[ 3 H]carnitine were incubated as described under "Experimental Procedures." Following the incubation, the sample was applied to Sephadex G-25 in.1 M potassium phosphate buffer, ph 7.4, at 4 C to remove DL-[ 3 H]carnitine which did not bind to CCBP. The pooled protein peak of this eluant then consisted of CCBP-carnitine complexes, and a sample of this eluant was assessed for the amount of radioactivity still bound. Samples of this eluant were treated in various ways as listed in Table V and the amount of radioactivity still bound to CCBP was measured using the micropartition systems described under TABLE IV Competition for CCBP binding to DL-[ 3 H]carnitine by structural analogs Duplicate assay tubes contained an analog at a concentration of 1 M, 5-1 Mg of CCBP, and DL-[ 3 H]carnitine at a concentration of 1 pm. Results are expressed as a percentage of the binding activity when no excess unlabeled analog was added. Analog Binding activity, Z zo 2Z None 1 L-Carnitine 37 D-Carnitine 87 y-aminobutyric acid 99 e-amino-n-caproic acid 99 5-Amino valeric acid 15 Lysine 78 Trimethyllysine N,N-Dimethylglycine Betaine 31 Choline 49 Acetylcarnitine 55 UNLABELED CARNITINE CONCENTRATION (AM) FIG. 3. Competition of CCBP binding tddl-[ 3 Hcarnitine by L-earnitine and D-carnitine. DL-[H]carnitine at a concentration of 1. /M was incubated with varying concentrations of L-carnitine (S) or D-carnitine (O). The binding activity was measured as described under "Experimental Procedures." The results have not been corrected for nonspecific binding.

5 Cardiac Carnitine Binding Protein 163 "Experimental Procedures." Table V suggests that the binding of DL-[3H]caritine to CCBP was reversible. Simply incubating the CCBP-carnitine complexes at 4 C resulted in dissociation of radioactivity. The dissociation of radioactivity was even more extensive with dialysis against potassium phosphate buffer or incubation with an excess of L-carnitine or an excess of D-carnitine or incubation with Triton X-1 at a concentration of.5%. Dialysis against mm DL-carnitine also resulted in extensive dissociation of radioactivity. The extent of the dissociation was greater with longer periods of dialysis. These results suggest that an exchange of unlabeled carnitine for the radioactive carnitine may have been occurring. Finally, a sample containing the CCBP-carnitine complexes which had incubated at 4 C for 24 h was incubated again with DL-[ 3 H]carnitine to measure the binding activity. This sample bound as much DL-[ 3 H]carnitine as that CCBP which had eluted from Sephadex G-25. Therefore, the binding capacity of CCBP was not affected by storage at 4 C for 24 h. Molstad et al. (19) have found a decreased uptake of carnitine by cultured myocytes in the presence of N-ethyhnlmaleimide, dinitrofluorobenzene, or 5,5'-dithiobis(2-nitrobenzoic acid) which was attributed to a participation of sulfhydryl groups in the transport process. To determine whether or not TABLE V Dissociation of radioactivity from CCBP Following the binding incubation, the sample was applied to Sephadex G-25 and the protein peak was collected and pooled. Aliquots were treated at 4 C as listed and then applied to an Amicon micropartition system as described under "Experimental Procedures" to assess the amount of radioactivity bound to CCBP. Results are expressed as the percentage of radioactivity still bound to CCBP relative to the amount of radioactivity bound to CCBP following elution from Sephadex G-25. Radioactivity Treatment steps remaining bound % Sephadex G-25 Incubation for 24 h 1 65 Dialysis against.1 M potassium phosphate buffer, 29 ph 7.4, for 24 h Incubate in 1 mm L-carnitine for 24 h 16 Incubate in mm D-carnitine for 24 h 23 Incubate with Triton X-1 for 24 h 32 Dialysis against 1 mm DL-carnitine for 2 h 35 Dialysis against 1 mm DL-carnitine for 18 h 16 z zo NEM CONCENTRATION FIG. 4. Effect of N-ethylmaleimide on CCBP binding activity. Aliquots of CCBP preparation containing 5-1 pg of protein were incubated with 5 nmol of DL-[H]carnitine and N-ethylmaleimide (NEM) at concentrations from.25 p[m to 2.5 mm in a final volume of.5 ml. Binding activity was measured and corrected fr nonspecific binding -s described under "Experimental Procedures." sulfhydryl groups are involved in binding carnitine to CCBP, several sulfhydryl group blocking agents were included in the incubation. Fig. 4 illustrates the dramatic decrease in binding of CCBP to carnitine in the presence of increasing concentrations of N-ethyhnlmaleimide. There was no binding of carnitine to CCBP at concentrations of N-ethylmaleimide greater than 1 M. Furthermore, when CCBP was treated with 1 mm N- ethylmaleimide and then dialyzed against.1 M potassium phosphate buffer, ph 7.4, to remove excess N-ethylmaleimide, the resulting preparation would bind less than 1.% of the carnitine that could be bound before treatment with N-ethylmaleimide. These data further suggest that N-ethylmaleimide modified sulfhydryl groups on CCBP necessary for binding activity. Iodoacetic acid eliminated 75% of the binding activity at 1 mrnm. Mercuric chloride completely inhibited the binding activity at concentrations greater than 1 M. These data indicate that sulfhydryl groups are involved in the binding of carnitine to CCBP. DISCUSSION The data reported demonstrate the existence of a cardiac carnitine binding protein. An in vitro binding assay was designed to allow characterization of the binding protein itself as well as characterization of the binding of carnitine to the binding protein. Optimal measurement of the binding activity involved incubating the protein preparation with DL-[ 3 H]carnitine for 6 min at 25 C. CCBP is a protein as judged by the presence of free sulfhydryl groups, by the heat lability of the binding component, and by the susceptibility to proteolytic action by pronase. Results from cell fractionation studies indicate that CCBP is associated with the plasma membrane of the cell which is consistent with a function of transport for CCBP and suggests that CCBP is the carrier for carnitine transport proposed by Molstad et al. (5). The data also distinguished CCBP from mitochondrial enzymes which utilize carnitine as a substrate, e.g. carnitine palmitoyltransferase, carnitine acetyltransferase, and carnitine translocase. The binding activity was specific for L-carnitine. D-Carnitine was less effective in competing for binding of DL-[3H]carnitine. The effective competition by y- butyrobetaine, choline, betaine, and N,N-dimethylglycine suggests that a methylated nitrogen was necessary for binding to occur. However, competition by N,N-dimethylglycine indicates that a trimethylated nitrogen was not an absolute requirement. Competition by choline indicates that a carboxyl group was not an absolute requirement for competition by the analogs. Competitive binding assays with L-carnitine and D- carnitine indicate that the affinity of CCBP for L-carnitine and for D-carnitine was quite different. D-Carnitine had little effect on the binding of DL-[ 3 H]carnitine at concentrations up to 4 M while L-carnitine competed with its maximum effect at a concentration of 1!LM. Molstad et al. (19) found that the cultured myocytes would take up D-carnitine as well as L- carnitine, but there was a greater affinity for the L-isomer. The binding of carnitine to CCBP was saturable, and a dissociation constant of.7 M was measured. Gel filtration chromatography of CCBP suggests that the binding protein aggregates to form very large complexes. This is supported by the decrease in binding activity apparent following centrifugation of binding protein samples from which Triton X-1 has been removed. This may also explain the apparent inhibition of binding activity by Triton since Triton would prevent aggregation of the binding protein which may be the form which is optimal for binding carnitine. The data suggest that CCBP may be associated with the plasma membrane of the heart cell, and we have hypothesized that CCBP functions as a part of the system to transport

6 164 carnitine into the heart cell. Thus, unless the carnitine-ccbp complex is internalized, the CCBP must release carnitine to the internal surface of the plasma membrane. Studies to look at the dissociative behavior of carnitine bound to protein indicate that the binding was reversible. Molstad et al. (19) found carnitine transport in cultured myocytes to be dependent on free sulfhydryl groups. Data presented here demonstrate that the binding of carnitine to CCBP was blocked when CCBP was modified with a sulfhydryl modification reagent such as N-ethylmaleimide, iodoacetic acid, or mercuric chloride. The characteristics described by Molstad et al. (5) for a protein carrier mediating the transport of carnitine into the heart cell were association of the protein with the plasma membrane, specific and saturable binding of carnitine, less effective competition for binding by D-carnitine than L-carnitine, competition of binding by structural analogs containing a trimethylamino group and a carboxyl group, and involvement of a critical sulfhydryl group in the binding. The characteristics of the CCBP fulfill these criteria for a protein carrier on the plasma membrane; thus, CCBP may mediate the transport of carnitine into the heart cell. Optimal function of normal cardiac energy metabolism would necessitate adequate carnitine transport mediated by a carrier protein which might be CCBP. Several biochemical lesions exhibit decreased levels of carnitine in the heart concomitant with a cardiomyopathy (2-23). Decreased cardiac carnitine levels may be partially responsible for the cardiomyopathy. The decreased carnitine levels in the heart may be due to altered binding activity of CCBP. Therefore, administering carnitine to a dystrophic patient with decreased skeletal muscle carnitine and impaired cardiac function might not be totally effective if the decreased tissue levels of carnitine were due to an impaired ability of the carrier protein to bind carnitine or to a decrease in the amount of binding protein. Further work needs to be done to purify CCBP to elucidate its role in the active transport of carnitine into cardiac tissue. Cardiac Carnitine Binding Protein REFERENCES 1. Broquist, H. P., and Borum, P. R. (1982) in Advances in Nutritional Research (Draper, H., ed) Vol. 4, pp , Plenum Press, New York 2. Neely, J. R., Rovetto, M. J., and Oram, J. F. (1972) Progr. Cardiovasc. Dis. 15, Haigler, H. T., and Broquist, H. P. (1974) Biochem. Biophys. Res. Commun. 56, Borum, P. R. (1978) Biochem. J. 176, Bohmer, T., Eiklid, K., and Jonsen, J. (1977) Biochim. Biophys. Acta 465, Vary, T. C., and Neely, J. R. (198) Fed. Proc. 4, Bahl, J., Vavin, T., Manian, A. A., and Bressler, R. (1981) Circ. Res. 48, Mlstad, P., Bhmer, T., and Hovig, T. (1978) Biochim. Biophys. Acta 512, Molstad, P., and Bhmer, T. (1981) Biochim. Biophys. Acta 641, Holloway, P. W. (1973) Anal. Biochem. 53, Bieber, L. L., Abraham, T., and Helmrath, T. (1972) Anal. Biochem. 5, Aronson, N. N., and Touster,. (1974) Methods Enzymol. 31, Lilienthal, J. L., Jr., Zierler, K. L., Folk, B. P., Buka, R., and Riley, M. J. (195) J. Biol. Chem. 182, Itzhaki, R. F., and Gill, D. M. (1964) Anal. Biochem. 9, Jones, L. R., Besch, H. R., Jr., Fleming, J. W., McConnaughey, M. M., and Watanabe, A. M. (1979) J. Biol. Chem. 254, Ramsay, R. R., and Tubbs, P. K. (1975) FEBS Lett. 54, Ramsay, R. R., and Tubbs, P. K. (1976) Eur. J. Biochem. 69, Pande, S. V., and Parvin, R. (1976) J. Biol. Chem. 251, Molstad, P., Bohmer, T., and Eiklid, K. (1977) Biochim. Biophys. Acta 471, Gilroy, J., and Meyer, J. S. (1975) Medical Neurology, pp , Macmillan, New York 21. Scott, R. S. (1975) Am. Heart J. 9, Feuvray, D., Idell-Wenger, I. A., and Neely, J. R. (1979) Circ. Res. 44, Shug, A. L., Thomsen, J. H., Felts, J. D., Bittar, N., Klein, M. I., Koke, J. R., and Huth, P. R. (1978) Arch. Biochem. Biophys. 187, 25-33