a Novel Radioactive Label for Determining Organ Sites of Catabolism of Proteins in the Circulation*

The primary tissue sites of catabolism of plasma proteins with long circulating half-lives are unknown. It has been difficult to identify these sites because plasma proteins are delivered to tissues at relatively slow rates but are rapidly degraded intracellularly within lysosomes. Therefore, a tracer attached to protein is lost from the site of uptake before an amount sufficient for quantitation can accumulate. We hypothesized that sucrose (Glupal-2 /3Fruf) would be a useful label to circumvent this difficulty because of the stability of sucrose in lysosomes; and thus, sucrose should remain in tissue long after the protein to which it was attached had been degraded to products released from the lysosome. [G-3HJRaffinose (RAF, GalpalGlupcwl2flFruf) was selected as the vehicle for attaching sucrose to protein. [31XjRAF was converted to the C-6 aldehydogalactose form with galactose oxidase and then covalently coupled to protein by reductive amination using NaBH3CN. [‘H]RAF was coupled first to two relatively long lived plasma proteins, bovine serum albumin and fetuin. The half-lives of these proteins in the rat circulation (&,* = -24 h) was unchanged, suggesting that RAF did not alter the normal mechanisms of protein clearance. When attached to short lived proteins with known sites of catabolism, such as asialofetuin, RNase B, and heatdenatured albumin, neither the tissue nor cellular sites of uptake of the proteins were altered. Thus, 13H]RAFasialofetuin was recovered almost exclusively (>90%) in the liver parenchymal cell fraction, while both [JHJRAF-labeled RNase B and heat-denatured albumin were recovered primarily (>85%) in nonparenchymal cells. In addition, the RAF label was observed to reside stably (tl,2 > 100 h) in the liver following degradation of the carrier protein; in contrast, radioactivity from “‘Ilabeled asialofetuin or RNase B was rapidly (tl,z < 30 min) lost from liver. Radioactivity from 13H]RAF-proteins was recovered in a lysosomally enriched subcellular fraction in liver and consisted of a low molecular weight species (-llOO), containing both glucose and fructose in a ratio similar to that in the original protein. The results of these studies establish that 13H]RAF

The primary tissue sites of catabolism of plasma proteins with long circulating half-lives are unknown. It has been difficult to identify these sites because plasma proteins are delivered to tissues at relatively slow rates but are rapidly degraded intracellularly within lysosomes. Therefore, a tracer attached to protein is lost from the site of uptake before an amount sufficient for quantitation can accumulate. We hypothesized that sucrose (Glupal-2 /3Fruf) would be a useful label to circumvent this difficulty because of the stability of sucrose in lysosomes; and thus, sucrose should remain in tissue long after the protein to which it was attached had been degraded to products released from the lysosome.
[G-3HJRaffinose (RAF, was selected as the vehicle for attaching sucrose to protein. [  (1). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 USC. Section 1734 solely to indicate this fact.
covalently coupled to plasma proteins should be a useful radioactive tracer for detecting the tissue and cellular sites of catabolism of long lived circulating proteins.
There is extensive information concerning the sites of synthesis and kinetics of turnover of plasma proteins; however, there is little firm evidence concerning their mechanisms of clearance from the circulation or their tissue sites of degradation (2)(3)(4). Perhaps the greatest technical difficulty in obtaining this information has been the absence of a suitable radioactive label with which to monitor the sites of catabolism of proteins with the long circulating half-lives characteristic of plasma proteins. In general, the rate of intracellular degradation of plasma proteins within lysosomes far exceeds the rate of clearance of these long lived proteins. Thus, if plasma proteins are labeled with conventional labels, such as ? or radioactive amino acids, the tracer does not accumulate at the site of catabolism, but, rather, as the protein is degraded to monomeric products which exit from the lysosome, the label is lost to cellular or extracellular metabolic pools, or both. In contrast, the sites of catabolism of proteins with short circulating half-lives are readily identified since their rates of clearance are relatively rapid, compared to the rate of lysosomal digestion. It has been possible, therefore, to identify the liver as the primary site for catabolism of ["H,%u]asialoceruloplasmin (5) and lz51-RNase B (6) injected in rats, and bovine /3-glucuronidase injected in mice (7), all proteins which exhibit half-lives of 15 min or less.
In order to circumvent the difficulties inherent in determining the in. uiuo fate of long lived plasma proteins, a tracer is needed which will 1) not affect mechanisms of protein clearance and catabolism, 2) be resistant to degradation by lysosomal hydrolases, and 3) accumulate in lysosomes at the site of the labeled protein's degradation. We hypothesized that sucrose (Glcpcyl-ZPFru,) covalently attached to protein through its glucosyl moiety might fulfill these criteria for the following reasons. First, because of the stable equilibration of inulin (a fructan) and rafiinose (RAF, Galpcul-GGlcpal-2/3Frun into plasma and extravascular spaces in uivo (8), it seemed unlikely that the terminal fructose residues on a protein would serve as recognition markers for a carbohydrate-dependent clearance process analogous to those already described for galactose (9), fucose (lo), mannose (11,12), and N-acetylglucosamine (13,14) terminal glycoproteins. Second, because of the absence of significant levels of mammalian lysosomal sucrase activity (15), it seemed unlikely that hydrolysis of the sucrose moiety would occur either before or after degradation of the protein. Finally, since sucrose does not readily diffuse from lysosomes (15), it or its amino acid conjugate should accumulate at this subcellular site of protein catabolism, despite the fact that other components of the protein would exit from the lysosome following degradation.
As a vehicle for the attachment of sucrose to protein, we chose the trisaccharide [G-3H]raffinose as our radioactive starting material. Oxidation of the C-6 hydroxyl of the galactose residue, using galactose oxidase (16), provided an aldehyde function which could be readily coupled to protein, under mild conditions, using NaBH3CN (17). We present below the results obtained in vitro for the coupling of RAF' to several proteins by this method. We also present our results from studies on the comparative fates of 12?-and ["HIRAFlabeled proteins in viuo, as a test of our hypothesis that RAF attached to protein would be a stable, inert, label useful for determining the organ and cellular sites of plasma protein catabolism.

In Vitro Experiments: Protein Modification
The overall scheme for the production of raffinaldehyde and its coupling to protein is shown in Fig. 1. As indicated, attachment of carbohydrate to protein by reductive amination required the presence of an aldehyde function on the sugar. This aldehyde function was introduced at the C-6 position of the galactose moiety of RAF using galactose oxidase. Catalase was included in the incubation mixture in order to destroy H202 produced during the oxidation reaction (16). Because of the selectivity of NaBH3CN for reduction of Schiff bases above pH 5 (25), the coupling reaction theoretically could be carried out simultaneously with the generation of aldehyde. However, since preliminary experiments indicated that NaBH3CN was inhibitory to galactose oxidase, the two reactions were carried out sequentially.
A pH of 7.5 was chosen as the pH for coupling in order to keep plasma proteins close to physiologic pH and yet have conditions alkaline enough to  Fig. 2. The reaction was found to proceed at a relatively rapid rate for approximately 4 h and then continued more slowly thereafter. It should be noted that the amount of protein-bound carbohydrate (acid-precipitable) was consistently found to be somewhat greater when determined by the anthrone method (21) than by radioactivity measurements.
The reported specific activity of the commercial ["H]RAF was determined and found to be correct, and its purity was confirmed by paper chromatography.
There was also no evidence of hydrolysis of [3H]RAF to its component sugars during the incubation with galactose oxidase or during the coupling reaction. The discrepancy between anthrone and radioactivity measurements may be explained by a kinetic isotope effect resulting from tritiation at the C-6 position of galactose and preferential oxidation of the unlabeled RAF by galactose oxidase. Neither increasing the time cf exposure of ["HIRAF to galactose oxidase, nor increasing the amount of enzyme in the standard incubations affected the extent of coupling of radioactivity to protein. Since radioactivity was the only parameter which could be conveniently assessed in either in uiuo experiments or when RAF was coupled to glycoproteins, the amount of RAF coupled was normally based on the specific activity of the [3H]RAF used. Under the standard conditions described in Fig. 2, 30 to 38 nmol of RAF were routinely coupled/100 nmol of albumin.  proteins tested (Experiments 9 to 11). As seen in Table I,  asialofetuin was a much poorer substrate for the coupling reaction than fetuin, itself; therefore, [3H]RAF-asialofetuin was prepared by enzymatic desialylation of ["HIRAF-fetuin. The differential reactivity of fetuin and asialofetuin in the coupling reaction may result from a conformational change in fetuin's structure upon removal of charged sialic acid residues. Alterations in other reaction parameters, particularly pH and time of incubation (26), can be expected to enhance the extent of RAF attached to proteins; in order to minimize structural alterations in proteins to be studied in uiuo particularly gentle conditions for coupling were chosen in these experiments.

In Viuo Experiments
Half-life Studies-Alterations in protein structure are often reflected by changes in their circulating half-life in uiuo (2). Therefore, by comparing the half-lives of l*'I-and [3H]RAFlabeled proteins in uiuo, the effect of RAF attachment on this biological property of proteins can be evaluated. The results of studies comparing the kinetics of clearance of several lz51and [3H]RAF-labeled proteins are shown in Fig. 3. Panel A demonstrates that, whether labeled with "'1 or ["HIRAF, both asialofetuin and RNase B were rapidly cleared from the circulation.
The rapid kinetics of clearance were expected since these proteins are recognized by separate carbohydratedependent clearance systems specific for glycoproteins with terminal galactose (asialofetuin (9)) and mannose (RNase B (11)) residues. The data of Panel A indicated that the attachment of RAF did not prolong the circulating half-life of asialofetuin and RNase B by interfering with these specific recognition systems. In Fig. 3B were essentially identical to that of '""I-fetuin, suggesting that the coupling of RAF did not cause significant changes in the physical or biological properties of the protein. Aa anticipated, the half-life of C3H]RAF-fetuin, -24 h, was in marked contrast to the 3-to 5-min half-life observed for the asialo derivative (9). Determination of the acid precipitability of 'H radioactivity recovered from plasma over a 72-h period yielded 93 to 99% precipitability.
In experiments with bovine serum albumin (Fig. 3C), one preparation was labeled with lz51, while another was first coupled with unlabeled RAF, in a molar ratio of 51 (RAF: albumin, determined by anthrone), and then iodinated by the lactoperoxidase technique (20). The circulating half-life of each preparation was -24 h. In data not shown, the half-life of a commercial preparation of L251-human serum albumin was 17 h. These data suggest that modification of albumin with up to 5 molecules of RAF did not cause either an increased or decreased rate of uptake of albumin from the circulation.
Tissue Studies--In order to verify the stability of r'H]RAF in tissue following protein uptake, the hepatic recovery of asialofetuin, heat-denatured albumin, and RNase B, labeled with either "'1 or [3H]RAF, was compared. Table II illustrates that at early times postinjection the extent of recovery of radioactivity from proteins labeled by either method was comparable.
However, for iodinated proteins, free iodide is rapidly released from the liver upon degradation of the protein (27). This loss is apparent even as early as 30 min, and by 24 h only traces of "'1 were recovered in liver from either lz51labeled asialofetuin or RNase B. In marked contrast, however, at 24    h postinjection, 90 and 74%, was matched by a concomitant appearance of 14 and 26% of dose, respectively, in urine at these two times. Radioactivity in lung, kidneys, spleen, and heart at these same time points accounted in toto for less than 1% of dose. These data demonstrate that the radioactivity lost from liver following uptake of cH]RAF-asialofetuin was not reutilized by other tissues but was excreted in the urine. In contrast to the gradual decline of recovered radioactivity following asialofetuin injections, radioactivity from ["HIRAFheat-denatured albumin or RNase B appeared to be more stable over a 48-to 72-h period. RNase B experiments were limited to 48 h because the animals had been nephrectomized. Since asialofetuin is known to be cleared in hepatocytes (9), whereas RNase B and heat-denatured albumin are cleared in hepatic non-parenchymal cells (2,6), differences observed in the stability of the RAF label from asialofetuin or heat-denatured albumin and RNase B injections may reflect differ-ences in the rate of turnover of cells or lysosomal contents in one cell type from another. In these experiments radioactivity not recovered in liver could be accounted for in blood and tissues of the reticuloendothelial system (see "Discussion"). The total body and hepatic distribution of radioactivity following inject,ion of ["H]RAF-RNase B was essentially identical to that previously reported for ""I-RNase B (6).

Cell
Separations-As a test of whether the RAF label would affect the known cellular sites of clearance of asialofetuin and heat-denatured albumin by parenchymal and nonparenchymal cells, respectively, the cell type in which radioactivity was localized after injection of each of these proteins was determined. Table III demonstrates that, based on the ratio of the specific activity (counts per min/cell) of the final purified parenchymal and non-parenchymal cell fractions, and an average ratio of two parenchymal cells/non-parenchymal cell in the initial cell suspension of liver (2&S), greater than 90% of the radioactivity from ["HIRAF-asialofetuin injections was accounted for in hepatocytes.
By a similar calculation, 85 to 90% of the hepatic radioactivity from heat-denatured albumin experiments was initially present in non-parenchymal cells. In experiments to be reported in detail elsewhere, ["HIRAF-RNase B was also recovered primarily (>85%) in non-parenchymal cells at 0.5 to 24 h postinjection.
There is some variation in relative specific activities of the two cell types among experiments, but this probably results from day to day variations in the cell isolation technique. The preferential localization of asialofetuin, heat-denatured albumin, and RNase B in the expected cell type supports the hypothesis that the attachment of RAF does not affect specific cellular uptake of proteins by pinocytic (asialofetuin, RNase B) or phagocytic (heat-denatured albumin) processes.

Subcellular
Localization-The expected lysosomal localization of recovered radioactivity from [3H]RAF-labeled proteins is illustrated in Table IV. The data also indicated that in addition to being stable in the liver for extended periods, the radioactivity from ["HIRAF protein injections remained in the lysosomally enriched subcellular fraction over a 48-to 72h period. The recovery of the lysosomal marker enzyme P-Nacetylhexosaminidase exactly paralleled that found for the radioactivity.
* The ratio of specific activity (sp. act.) is based on the counts per min/cell of the purified parenchymal (PC) and nonparenchymal (NPC) cells obtained from livers at indicated times after collagenase perfusion and differential centrifugation as described under "Experimental Procedures." Average purities of each cell type were 96 to 99% and average recoveries of parenchymal and non-parenchymal cells were 70 and 30% of cells in the initial cell suspension of liver, respectively.
' Calculated per cent of radioactivity in parenchymal or non-parenchymal cells in whole liver based on ratio of specific activities and observed average ratio of parenchymal/nonparenchymal cells of 2.  Fig. 4B, the molecular weight of the unknown peak was approximately 1100. Since essentially 100% of dose from asialofetuin injections was accounted for by the sum of recovered radioactivity in liver and urine, it seems likely that the entire RAF molecule is recovered attached to a small peptide derived from the protein. Paper chromatography of hydrolysates of the degradation products recovered from the P-2 column indicated that both radioactive fructose and glucose are recovered in a ratio essentially identical to that found for the starting protein. DISCUSSION A major gap in our knowledge of the metabolism of plasma proteins is the lack of essential information on the regulation of homeostatic concentrations of these proteins in plasma. To understand how plasma protein metabolism is regulated, it is at least necessary to identify the sites of catabolism of these proteins. The experiments described in this report represent an effort to overcome a major obstacle in the accurate determination of these sites of catabolism, namely that, although these proteins are delivered to tissues at a relatively slow rate (e.g. tlrz r 24 h), rates of degradation in tissues (e.g. tlr2 5 15 min) are relatively rapid. Thus, any radioactive tracer attached to protein is lost from the site of uptake at a rate too great to allow quantitation.
We sought to develop an inert, indigestible radioactive tag which would accumulate at the tissue site of a protein's degradation and, thus, measure the cumulative degradation of the protein in that tissue. Sucrose seemed especially attractive as a radioactive ligand for this purpose because of its demonstrated stability in lysosomes. Recently, based on a similar rationale, Pittman and Steinberg (29) reported the attachment of [14C]sucrose to low density lipoprotein and albumin, via a succinate spacer arm. Their in vitro studies of the uptake of the labeled proteins by cells in culture demonstrated both accumulation and stability of the radioactive label in cells over a 6-h period. While these initial observations are promising, there are limitations inherent in this method for attaching sucrose to protein. The attachment through the succinate spacer alters the charge of protein both by conversion of an amine to an amide function and by addition of negative charge from uncoupled succinate groups on the sucrose moiety. In addition, the ester linkage between sucrose and the linking succinate arm may well be susceptible to plasma esterases in viva In contrast, use of reductive amination to couple RAF, and thus sucrose, to protein results in minimal alterations in charge and isoelectric point of the modified protein. Further, we have observed that the label is completely stable under in viuo conditions. As noted above, the coupling procedure can be performed under mild, physiological conditions and does not result in any modification of gel chromatographic behavior of the proteins studied or the enzymatic activity of RNase B. The studies presented also document that the kinetics and sites of clearance of RAFlabeled proteins are essentially identical to those found for '?-labeled proteins, suggesting that, in fact, there is no major alteration in the physical and biological properties of the modified proteins.
A basic assumption in the choice of sucrose as in indigestible tracer is the absence of sucrase or ,&fructofuranosidase activ-ities in mammalian lysosomes. There has been, however, a report of a low but detectable level of sucrase in rat liver lysosomes (30). As shown above, while the RAF label is particularly stable in non-parenchymal cells, there was a measurable loss of the RAF radioactivity with time from hepatocytes. This could result from normal hepatocyte turnover but is unlikely to result from enzymatic degradation of the RAF-amino acid conjugate. As noted above, 100% of injected dose of ["H]RAF-asialofetuin could be accounted for in liver and urine at all times. In addition, only a single low molecular weight species was detected on Bio-Gel P-2 from either of these sources, for as long as 96 h postinjection, and contained both glucose and fructose in the expected ratios. These data suggest that either RAF attached to protein is not a suitable substrate for the sucrase described by Horvat (30) or that the sucrase activity is too low to be detectable during the course of our experiments.
It should also be noted that, since the loss of radioactivity from hepatocytes is not detectable for at least 24 h postinjection and has an approximate half-life of 100 h, experiments to study the sites of uptake of proteins with long circulating half-lives can be carried out over at least a 48-h period, with little loss of accumulating radioactivity in parenchymal cells, should this prove to be a site of uptake.
The emphasis of the experiments reported here has been to study the stability of the [3H]RAF label in liver, as a model for its stability in other tissues. The data obtained to date verify this stability in liver and in the lysosomal subcellular fraction of both hepatic cell types. In preliminary studies, radioactivity not found in liver from [3H]RAF-heat-denatured albumin injections could be accounted for primarily in kidney, spleen, and lung, other organs of the reticuloendothelial system expected to participate in the clearance of aggregated protein (2). The recovery of radioactivity in these organs appears to have a stability comparable to that found for liver, and only 6.5% of dose was found in urine over a 72-h period. The uptake of RNase B also occurs in organs of the reticuloendothelial system, but a large proportion is found in muscle and hide (6). Thus, this protein should also be useful for evaluating the stability of RAF in nonhepatic tissue. Since the ultimate goal of our research is to investigate regulation of plasma protein catabolism, it was essential to document that the RAF label did not interfere with normal mechanisms of clearance or elicit clearance by different pathways. The recovery of asialofetuin in hepatocytes and RNase B and heat-denatured albumin in non-parenchymal cells of liver suggests that the normal mechanisms of clearance were operating on RAF-labeled proteins. The fact that ?-and RAF-labeled fetuin and albumin were cleared from the circulation with similar kinetics also suggests that attachment of RAF did not interfere with the normal uptake mechanisms for these proteins.
One other important aspect of the use of RAF as a label is that it should accumulate with time in the tissues active in catabolism of long lived proteins, as protein is slowly delivered to those tissues. In one set of preliminary experiments studying the hepatic recovery of [3H]RAF-fetuin (circulating halflife, -24 h), the recovery of radioactivity at 12, 24, and 48 h postinjection was 21, 29, and 36% of dose, respectively. At these same times there was a concomitant decrease in radioactivity in plasma from 41 to 23 to 12% of dose (Fig. 3). These encouraging initial findings with [3H]RAF-fetuin are entirely consistent with the hypothesis that the recovery of the RAF label from a protein with a long plasma half-life would show a progressive increase in a tissue involved in the catabolism of that protein, as it is slowly removed from the circulation. An additional 7 and 14% of dose from C3H]RAF-fetuin was re-covered in urine at 24 and 48 h postinjection. Thus, blood, liver, and urine accounted for about 60% of dose at the three times studied. The remainder of radioactivity could be accounted for in other visceral organs and extravascular compartments.
Details of these experiments will be presented elsewhere.
In summary, from the experiments presented above, we conclude that [3H]RAF functions as a biologically inert, indigestible tracer which accumulates in tissues following protein removal from the circulation and intracellular degradation in lysosomes. We believe that the technology which we have developed here provides an exciting new tool for investigating the sites, and eventually, the regulation of plasma protein catabolism in uiuo.