Partial Purification and Characterization of a Binding Protein for Biologically Active Phorbol and Ingenol Esters from Murine Sera*

We have purified a protein (Mr - 71,000) from murine sera 104-fold which directly binds biologically active phorbol esters, ingenol esters, and mezerein in a specific, reversible, and saturable manner. The binding of labeled phorbol-12,13-dibutyate (PDBu) to the purified protein is rapid and dose-dependent. Those phorbol and ingenol esters which stimulate cell growth in culture and have tumor-promoting activity in vivo inhibit the binding of labeled PDBu, while the biologically inactive derivatives fail to do so. Other nonditerpene tumor promoters, retinoids, steroids, and prostaglan-dins do not interfere with PDBu-protein interaction. Epidermal growth factor, insulin, bovine serum albumin, hemoglobin, ovalbumin, ferritin, myoglobin, fe- tuin, and lipase do not interact directly with PDBu. The purified binding protein competitively inhibits the binding of PDBu to its specific receptors. It is nongly-cosylated and slightly hydrophobic. The protein is heat- and acid-labile and is present in sera of various mammalian species. Its concentration in murine sera is age-, sex-, and strain-independent.

Tumor promoters are compounds which are themselves noncarcinogenic but which can induce tumors in animals previously treated with a suboptimal dose of certain chemical carcinogens (1-5). Most of the experimental work on tumor promotion has been carried out with phorbol esters, especially 12-O-tetradecanoylphorbol-13-acetate, initially isolated from croton oil derived from the seed of the plant Croton tiglium (4,6). TPA' and other biologically active phorbol esters elicit and modulate a variety of biochemical and biological responses in mouse skin, including stimulation of macromolecular synthesis, histone phosphorylation, synthesis of phospholipids, and modulation of the metabolism of polyamines and cyclic nucleotides (1-5, 7-13). In addition, these compounds induce ultrastructural changes in and affect the differentiation of murine epidermis (7, 14). Tumor-promoting phorbol esters also evoke pleiotypic responses in cultured cells, including the stimulation of macromolecular synthesis and cell proliferation, induction of plasminogen activator and ornithine decarboxylase, loss of surface-associated fibronectin, alterations in the metabolism of cyclic nucleotides and polyamines, stimulation of prostaglandin synthesis, either the inhibition or stimulation of differentiation, and alterations in cell morphology and cell permeability, and elevation in the level of (Na', K+)-ATPase activity (1-5, 15-23).
Several biochemical and biological studies provide evidence that the initial site of action of tumor-promoting phorbol esters may be the membrane of target cells (3-5, 21, 23-26). The tumor-promoting phorbol esters have been found to modulate the interaction between epidermal growth factor and its membrane receptors in a variety of cells in culture (27-31). The pleiotypic effects of TPA and reiated tumor promoters in vivo as well as in vitro seem to mimic the several actions of growth-stimulating polypeptide hormones such as EGF (32) and sarcoma growth factor (33). However, the~effect, although rapid in modulating the EGF receptors, is indirect as it cannot be shown using low temperatures (28) and/or fmed cells or in isolated cell membranes.' This would suggest that TPA produces its membrane effects through an interaction distinct from the EGF-receptor interaction.
Recently, we and others have reported the presence of specific high affinity receptors for biologically active phorbol and ingenol esters in a variety of cells and tissues using [3H] phorbol dibutyrate as a ligand (34, 35). The discovery of specific receptors for biologically active phorbol and ingenol esters, compounds of plant origin, led us to propose that TPA and certain analogues may have some structural resemblance to the endogenous growth promoting and/or differentiation modulating substance(s) (agonists or antagonists) that have specific membrane receptors. These compounds recognize and interact with the receptors, mimicking the action of the endogenous putative ligand(s) (34).
During the extensive search for the putative endogenous ligand(s) for the receptor, we have found that sera from a variety of mammalian species contain a protein which competitively inhibits the binding of PDBu to its receptors. Further investigation has revealed that this protein is a binding protein for biologically active phorbol and ingenol esters. We report here the partial purification and characterization of the binding protein from murine sera.

Phorbol
Ester-binding Protein subsequent steps were done at 0-4 "C. The homogenate was centrifuged for 10 min at 1700 X g, the supernatant was removed and centrifuged for 60 min at 105,000 X g. The resulting pellet was suspended in one-fourth of the initial volume of PBS. The suspension was aliquoted into small volumes and stored at -70 "C.
PDBu-binding Assays-The binding of [3H]PDBu to cells was performed as previously described (34). The binding of [3H]PDBu to BMF or soluble receptors was performed in duplicate in disposable glass tubes (12 X 75 mm) (Kimax) either in the absence or presence of 20 pg/ml of unlabeled PDBu. The binding mixture contained 5 ng of [3H]PDBu (-4 X lo4 cpm), 0.2% final concentration of dimethyl sulfoxide and BMF (-50 pg of protein) or soluble receptors in a total volume of 0.25 ml of binding buffer consisting of Dulbecco's minimum essential medium (DME medium) containing bovine serum albumin (1 mg/ml) and N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (5 mM), adjusted to pH 6.8. After incubation for 30 min at 23 "C, the tubes were chilled, 0.25 ml of cold 4% calf serum (Colorado Serum Co., Denver, CO) and 0.7 ml of cold 25% polyethylene glycol 6ooo in 1 mM Tris-HC1, pH 7.4, was added to each tube, the contents vortexed, the tubes allowed to stand for 15 min at 4 "C, and centrifuged for 10 min at 1500 X g at 4 "C. The supernatant solution was carefully drained off, the pellet was suspended in 2.5 ml of cold 10% PEG in 0.1 M Tris-HC1, pH 7.4, tubes were again centrifuged as earlier, supernatant solution was removed, and pellets were solubilized in 0.8 ml of lysing buffer (0.01 M Tris-HC1, pH 7.4, containing 0.5% sodium dodecyl sulfate and 1 mM EDTA). The mixture was transferred to counting vials and 10 ml of Aquassure (NEN) was added to each vial. The vials were vigorously shaken and radioactivity determined using a Beckman p counter. The radioactivity bound in the presence of 20 pg/ml (50 to 200 cpm) of unlabeled PDBu was considered to be nonspecific and all data were corrected accordingly.
Assays for PDBu Zinding Inhibiting Activity-The competition assays either with cells or with BMF were performed as described above except the test material was added to the reaction mixture along with other components.
Gel Filtration Assay for PDBu Binding-The reaction mixture consisting of 2.5 ng of [3H]PDBu (-20,000 cpm), 0.5% f'mal concentration of Me2S0, 0 or 5 pg of unlabeled PDBu and soluble receptors or binding protein in a total volume of 0.125 ml was incubated for the desired time and temperature. The mixture was applied to a column of Bio-Gel P-10 (0.9 X 7.9 cm) previously equilibrated with phosphatebuffered saline (pH 7.1) at 4 "C. The flow rate was maintained at approximately 10 ml/h, IO-drop fractions were collected and radioactivity determined as described above.
Protein Estimation-Protein concentrations were determined by the method of Lowry et al. using bovine serum albumin as the standard (37).
Purification of PDBu-binding Protein-Forty ml of murine sera were applied to a column of Sephadex G-200 (5 X 90 cm) previously equilibrated with PBS at 4 "C. The flow rate was approximately 30 ml/h and 300-drop fractions (20 ml) were collected. Fractions 50 to 57 were pooled ( Fig. 1) and applied to a column of phenyl-Sepharose CL-4B (1.5 X 7 cm) previously equilibrated with PBS at room temperature. The column was washed with approximately 100 ml of PBS. The bound proteins were eluted with distilled water. Five-ml fractions were collected. An aliquot of each fraction was assayed for PDBu binding inhibiting activity. Most of the activity appeared in fraction 3 ( Fig. 2) which was divided into 0.5-ml aliquots, lyophilized, and stored at -70 "C. Each aliquot was dissolved in 0.5 ml of cold PBS for use in these studies.

RESULTS
Purification of the Protein-The elution profiie of the PDBu binding inhibiting activity of murine sera from a column of Sephadex G-200 is shown in Fig. 1. The inhibiting activity emerged from the column as a single peak with a median size slightly larger than bovine serum albumin (Mr -71,000). Fig. 2 depicts the results of chromatography of pooled Sephadex G-200 fractions containing the inhibiting activity on phenyl-Sepharose. Most of the activity was eluted by water in fraction 3. The purification is summarized in Table I. A 104fold purification with a 57.7% yield has been achieved in a two-step process. Attempts to purify the protein further by ion exchange chromatography or by affinity chromatography on various kinds of lectin-Sepharose have not been successful.
The competition of binding of labeled PDBu to BMF by various fractions (Table I) is presented in Fig. 3. The degrees of inhibition increased with increasing concentrations of protein. However, the maximum inhibition by unfractionated sera was never found to be more than 58% even at very high concentrations of serum (20 mg of protein in 0.25 ml of total reaction mixture).
The partially purified PDBu binding inhibiting activity is heat-and acid-labile, as well as protease sensitive. However, PBIA is resistant to DNase, RNase, neuraminidase, or galactosidase. Furthermore, protein is not retained on various lectin-Sepharose columns indicating the absence of glycosyl residues. The activity is stable to extensive dialysis against water (4 "C), lyophilization, and storage at -70 "C (-6 months with no appreciable loss of activity).
Distribution of Protein (PBIA) in Sera from Various Species-We tested mouse, rat, hamster, guinea pig, rabbit, goat, fetal calf, calf, monkey, and human sera for the presence of PBIA in order to determine the richest source for purification purposes ( Table 11). The activity was found to be highest in     Fig. 4. The inhibitory effects of the protein were much greater at lower concentrations of PDBu. As the concentration of PDBu was increased, decreasing the ratio of receptors to PDBu molecules, the protein-elicited inhibitory effects lessened until they vanished at a PDBu concentration   Fig. 4 using a Scatchard plot revealed the decrease of receptor affinity in the presence of the protein (Fig. 4B). Fig. 5 shows the effects of protein concentration on the extent of PDBu binding to BMF in the absence and presence of protein at two different concentrations. The inhibition was not overcome by increasing the receptor concentration, again indicating the competitive nature of inhibition.
Direct Binding of PDBu to Protein-PBIA could act to inhibit the binding of PDBu to receptors using one of the following mechanisms: 1) directly or indirectly masking or destroying the receptors, 2) transforming PDBu to a form incapable of binding to receptors, or 3) directly binding to PDBu or masking PDBu. We performed experiments to test these possibilities and found that PBIA binds directly to biologically active phorbol and ingenol esters. Fig. 6 shows the direct binding of [3H]PDBu to the partially purified protein. Free [3H]PDBu was not found to be excluded from Bio-Gel P-10 and most of the radioactivity appeared in ever, the distribution of label in different fractions at the lower concentration of protein was unexpected and intriguing. As the concentration of protein was increased, the radioactivity started to shift from the unbound position (A) toward void volume fractions and finally co-eluted with the protein ( F ) .
We expected that [3H]PDBu would either appear in fractions 3 to 5 (bound position) or fractions 7 to 9 (unbound position), and that the amount of label in fractions 3 to 5 would be directly proportional to protein concentration similar to the binding of [3H]PDBu to soluble brain receptors (as shown in Fig. 7b). We feel that the on and off rates of PDBu binding to protein depend on protein concentration and at lower concentrations of protein, PDBu is rapidly dissociated and during filtration does not have an opportunity to recombine with protein and co-elute with the protein.
Protein bound very rapidly to PDBu. The binding was reversible. When an aliquot of fraction 4 ([3H]PDBu-BP complex) was incubated for 10 min at 4 or 23 "C and rechromatographed, almost all the radioactivity appeared at the position of free [3H]PDBu.  (1-5).

Binding Protein Does not Alter PDBu-['H]PDBu was
incubated with binding protein. An aliquot was passed over a Bio-Gel P-10 column. Bound radioactivity and total radioactivity was extracted with a mixture of chloroform and ethanol (2:l) after adding carrier PDBu. The organic phase was recovered, dried under nitrogen, and suspended in a small volume of MeZSO. The concentrated materials were analyzed by high performance liquid chromatography using a C18-Bondapak/Porasil column (Waters Associates). All the radioactivity coeluted with unlabeled PDBu from both extracts. No other radioactive peaks were found. Thus, binding protein does not alter PDBu during the interaction.

DISCUSSION
Our results demonstrate the existence of a protein (approximate M, 71,000) in murine and other mammalian sera which specifically interacts with biologically active phorbol and ingenol esters, and mezerein but not with biologically inactive derivatives of these diterpenes. The binding protein has been partidy purified from murine sera. PDBu interaction with this protein is specific, reversible, and saturable. The nonditerpene ester tumor-promoting agents apparently do not interact with this protein, consequently, they do not modulate PDBu-BP interaction. Also, potent inhibitors of tumor promotion such as retinoids, anti-inflammatory steroids, prostaglandins, and cyclic nucleotides fail to affect the binding of PDBu to the protein. The above characteristics of the binding proteins are quite similar to the interaction between PDBu and its receptors in cells and tissues (34). However, BP appears to be different from PDBu receptors in many respects. The binding of PDBu to protein appears to be time-and temperature-independent and very much protein concentration-dependent (Fig. 7). In contrast, PDBu binding to its specific receptors is a concentration-, time-, and temperature-dependent phenomenon (34). The on and off rates of PDBu-BP interactions are apparently very different from PDBu-receptor interaction. Soluble brain receptors for PDBu resolve into three distinct peaks (Mr -440,000,230,000, and 94,000) during gel Titration (Bio-Gel A-0.5)2 while binding protein appears as a single peak (Mr -71,000) (Fig. 1). Finally, binding protein inhibits the binding of labeled PDBu to cells or BMF, whereas soluble brain receptors show synergistic effects on the binding of PDBu to BMF. However, the final answer to the question of chemical relatedness and differences between the binding protein and receptors would come after purification to homogeneity of these proteins and their physical and chemical analyses.
Although the binding protein competitively inhibits the binding of PDBu to its receptors in cells or to BMF, it is not an endogenous ligand (either agonist or antagonist) (34) of PDBu receptors. It neither modulates EGF receptor interaction nor affects TPA-elicited modulation of EGF binding to its receptor (28). This protein neither induces nor modulates TPA-induced adhesion of human promyelogenous leukemia cells HL60 (38)." We routinely use these two tests in our laboratory to determine the biological responses of tumorpromoting diterpene esters. Although the temptation is great, a competitive inhibitor of animal origin of PDBu binding to its receptor should not be christened an endogenous ligand unless it elicits biological responses akin to TPA in vivo and in vitro or it modulates TPA-induced biological responses.
Thus, why should mammalian sera contain specific binding protein for biologically active phorbol and ingenol esters, compounds of plant origin? The discovery of binding protein strengthens our belief in the proposal we presented in the report on specific receptors for these compounds (34). We now propose that TPA and certain analogues may have a structural resemblance to the endogenous growth-promoting and/or differentiation-modulating substance(s) (agonists and/or antagonists) that have specific binding proteins and specific receptors. These compounds recognize and interact with binding proteins and receptors, mimicking the action of the putative substances. Azaserine, cordycepin, curare, opiate, physostigmine, plant lectins, puromycin, and tubercidin appear to exert their action by such biological mimickry (39-41). The isolation and characterization of putative endogenous ligand(s) should help in understanding the mechanism of tumor-promotion.