Lead-binding Properties of Intestinal Calcium-binding Proteins*

The bovine and chick vitamin D-induced intestinal calcium-binding proteins (CaBP) bind lead. Bovine CaBP binds 2 atoms of lead/molecule, and chick CaBP binds 4 atoms of lead per molecule and these values are identical to those for calcium binding. 46Calcium-dis- placement studies indicate significantly higher affinities for lead than for calcium for both proteins. All evidence indicates that lead is bound to the 4 high affinity calcium-binding sites on chick CaBP and to the corresponding 2 high affinity sites on bovine CaBP, and that binding of lead to sulfhydryl groups is, relatively, not significant. Calmodulin, troponin C, and oncomodulin also bind lead with high affinities and in preference to calcium, indicating that lead binding is a general property of proteins belonging to the troponin C superfamily of calcium-binding proteins. The vitamin D-induced intestinal calcium-binding proteins (CaBP’) represent the best defined molecular expression of the vitamin D endocrine system (1). Whereas these proteins bear significant relationship to intestinal calcium absorption, their precise function in this process remains unknown. The amino acid sequence of the bovine CaBP ( M = 8500) is known (2), as is the crystal structure (3) and the solution conformation (4). This protein binds 2 Ca(I1) atoms/molecule with an apparent intrinsic association constant (kaC“) of 4.3 X lo6 M” (5). Two sequence regions have tentatively been designated as the calcium-binding domains on the basis of the E-F hand concept (2, 3). In addition,

The bovine and chick vitamin D-induced intestinal calcium-binding proteins (CaBP) bind lead. Bovine CaBP binds 2 atoms of lead/molecule, and chick CaBP binds 4 atoms of lead per molecule and these values are identical to those for calcium binding. 46Calcium-displacement studies indicate significantly higher affinities for lead than for calcium for both proteins. All evidence indicates that lead is bound to the 4 high affinity calcium-binding sites on chick CaBP and to the corresponding 2 high affinity sites on bovine CaBP, and that binding of lead to sulfhydryl groups is, relatively, not significant. Calmodulin, troponin C, and oncomodulin also bind lead with high affinities and in preference to calcium, indicating that lead binding is a general property of proteins belonging to the troponin C superfamily of calcium-binding proteins.
The vitamin D-induced intestinal calcium-binding proteins (CaBP') represent the best defined molecular expression of the vitamin D endocrine system (1). Whereas these proteins bear significant relationship to intestinal calcium absorption, their precise function in this process remains unknown.
The amino acid sequence of the bovine CaBP ( M = 8500) is known (2), as is the crystal structure (3) and the solution conformation (4). This protein binds 2 Ca(I1) atoms/molecule with an apparent intrinsic association constant (kaC") of 4.3 X lo6 M" ( 5 ) . Two sequence regions have tentatively been designated as the calcium-binding domains on the basis of the E-F hand concept (2, 3). In addition, structural changes induced by the binding of calcium result in a molecule which is functionally resistant to tryptic digestion (6). The chick CaBP (Mr = 28000) binds 4 Ca(I1) atoms/molecule with high affinity (kaCe = 2 X lo6 M-') (7). Both CaBPs bind several other cations, notably Sr(II), Ba(II), Cd(II), and the La(II1) series elements (8,9), in a fashion apparently related to their ionic radii relative to that of Ca(I1).
Recent reports linking intestinal lead absorption to dietary vitamin D, calcium, and phosphate status (10, ll), as well as * This work was supported by Contract EV-S-2792 from the Department of Energy and Grant AM-04652 from the National Institutes of Health. 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 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. § On sabbatical leave from the Biochemistry Department, the Weizmann Institute of Science, Rehovot, Israel, 76100.
The abbreviations used are: CaBP, calcium-binding proteins; ka, apparent intrinsic association constant; (Ca)F, (Pb),, free concentrations of Ca(I1) and Pb(II), respectively; a, fractional inhibition of binding; U, average number of Ca(I1) or Pb(I1) bound to the number of high affinity binding sites; Pipes, piperazine-N,N"bis(2-ethanesulfonic acid); EGTA, ethylene glycol bis(p-aminoethyl ether)-N,N,N',N"tetraacetic acid. the demonstration of a close relationship between lead absorption and intestinal CaBP content (11, 12), suggest the possible involvement of CaBP in intestinal lead transport.
The possibility of such involvement, together with the observation that rat intestinal CaBP binds lead (13), prompted the present studies in which we examine, in some detail, the lead-binding properties of the bovine and chick CaBPs, as well as several other high affinity Ca(I1)-binding proteins.

EXPERIMENTAL PROCEDURES
The vitamin D-induced chick and bovine CaBPs were purified as described previously (14). Bovine brain calmodulin was provided by Dr. Barry Levine, Oxford, England, bovine skeletal troponin C, by Dr. James Potter, University of Miami, Florida, and oncomodulin, by Dr. John MacManus, National Research Council, Canada. All proteins were subjected to a preliminary dialysis versus 3 X 1000 volumes of 1.0 mM Pipes, 150 mM KC1 buffer, pH 6.5, at 4 "C for 48 h. Equilibrium dialysis experiments were carried out in this same buffer containing added lead or calcium at 4 "C for 48 h with moderate stirring. Dialysis sacs were prepared from Spectrapore 1 membrane tubing (National Scientific, Cleveland, Ohio) with a nominal M, cut off of 8000. There was no detectable leakage of the proteins through this casing.
Calcium and lead concentrations were determined directly by atomic absorption spectrophotometry (Model 360 AAS, Perkin-Elmer, Norwalk, CT) where concentrations of these elements permitted, or by specific activity measurements, where necessary. Stock lead and calcium solutions were prepared from analytical grade reagents and the concentrations determined by atomic absorption spectrophotometry relative to commercially prepared standards (Fisher).
"Calcium and *031ead were purchased from New England Nuclear. Liquid scintillation counting and y spectrometry were performed using a Beckman LS 200 and Beckman 7-300 counter, respectively. Protein concentrations were determined by amino acid compositional analysis on a Beckman 119C1 analyzer, as previously described (2).
Cysteine groups on the chick CaBP were blocked by S-pyridylethylation following a modification of the method of Friedman et al. (15). Chick CaBP (2.5 mg lyophilized from water) was dissolved in 0.5 ml of 2 M N-ethylmorpholine acetate buffer, pH 8.0. One pl of 0mercaptoethanol was added and incubated at room temperature for 2 h, followed by 3 pl of 4-vinyl pyridine (Aldrich) for 90 min. S-Pyridylethylated CaBP was separated from reaction products on a 0.9 cm X 25-cm column of Sephadex G-25 (fine) equilibrated with 1.0 mM Pipes, 150 mM KC1 buffer, pH 6.5, and stored frozen until use.

RESULTS AND DISCUSSION
Direct analysis of lead binding to chick CaBP was first accomplished by equilibrium dialysis, using '03Pb and Scatchard plot data representation. Chick CaBP (1. CaBP with high affinity (haPb = 1.1 X lo6 M-') and 1 or 2 additional moles of Pb(I1) with hapb = 7 X 10' "I.
Since ambient Ca(I1) concentrations, arising from the dialysis buffer and the chick CaBP, were determined to be relatively high (6.25 X M) at all concentrations of Pb(II), competition for binding sites between Pb(I1) and Ca(1I) could be significant. At this free Ca(I1) concentration, the computed occupancy of Ca(I1)-binding sites, by Ca(I1) (in the absence of Pb(I1) and assuming a k, for Ca(I1) of 2 X lo6 "I) would be 3.7, effectively diminishing the experimentally determined apparent haPb considerably.
Straightforward determination of kapb by simple Scatchard plot representation was therefore precluded. Similarly, methodological difficulties obviated the use of EDTA-or EGTA-Ca(I1) or -Pb(II) buffer systems in the presence of 2 soluble binders and 2 cations. As an approach to estimation of kaPb, Displacement of bound Pb(I1) by Ca(I1) was next examined for chick CaBP. Following sampling for the 203Pb-binding data, stable Ca(1I) was added to each outside solution to a final concentration of 1 x M (equivalent to the highest Pb(I1) concentration). Dialysis proceeded for an additional 48 h and 203Pb binding was again determined. The results (Fig.  1) clearly show '03Pb displacement, in all cases, by added Ca(I1). In addition, while the entire binding curve is shifted, reflecting lower occupancy of sites by Pb(II), the apparent k, " (-slope) for both the high and lower affinity sets of sites remain essentially unaltered, confirming the competitive nature of Ca(I1) and Pb(I1) binding.
Wheres estimation of kaPb is possible at single concentrations of Ca(I1) and Pb(I1) (above), corroboration of this value by multiple point analysis via a Ca(I1)-displacement curve was accomplished for both the bovine and chick CaBPs. The displacement of 45Ca by Pb(I1) at differing Pb(I1) concentrations (from 5 X M to 1 X M) at a single stable Ca(I1) concentration (1.1 X M for bovine and 1.5 X M for chick CaBP) was examined by equilibrium dialysis. In all cases, CaBPs were pre-loaded with Ca(1I) and 45Ca prior to addition of Pb(I1).
The plotted data (Fig. 2) show the Pb(I1) concentrations necessary to displace 50% of the bound Ca(II), under conditions described above, to be 1.8 X M for the chick and 4.3 X 1O"j M for the bovine CaBPs. At the known concentration of Pb(I1) sufficient to displace 50% of the bound Ca(II), Equation 1 becomes simplified to the following.
Computation of kaPb for the chick and bovine CaBPs yield values of 1.61 X lo7 M-' and 1.08 X lo7 M-', respectively, ionfirming that the actual affinities of both proteins are greater for Pb(I1) than Ca(I1).
Ca(I1)-and Pb(I1)-binding to the chick and bovine CaBPs were examined by equilibrium dialysis under several sets of conditions (Table I)   The native bovine CaBP was also shown to bind 2 mol of Ca(I1) or 2 mol of Pb(II)/mol of protein with a combined total binding capacity for Ca(I1) and Pb(I1) of 2 mol/mol of protein.
The implication of experiments thus far, based on similar stoichiometry of binding and mutual displacement, was that Pb(I1) and Ca(I1) compete for binding at the same set(s) of sites. However, the proclivity of lead to interact with sulfhydryl groups to form lead mercaptides raised the possibility of Pb(I1) binding to protein SH groups (secondary sites) with concomitant stoichiometric allosteric release of Ca(I1) from the Ca(I1)-binding sites.
The significance of Pb(I1)-sulfhydryl interaction was examined by equilibrium dialysis in the presence of 1 mM 0mercaptoethanol (13-fold molar excess over protein sulfhydryls) and equimolar concentrations of Pb(I1) and Ca(II), as well as Ca(I1) alone. Whereas Ca(I1) binding to CaBP was not influenced by P-mercaptoethanol (Table I), Pb(I1) binding was diminished somewhat. This reduction was not significant in terms of molar ratio of protein SH and 8-mercaptoethanol, however, and may well have been due to Pb(I1)-mercaptoethanol interaction, effectively reducing the free Pb(1I) concentration. The involvement of protein sulfhydryl groups in Pb(I1)binding to CaBP was essentially eliminated by forming the S-8-pyridylethylcysteine derivative of chick CaBP and repeating the dialysis experiments. Pb(I1) binding to the SHblocked protein (100% of total cysteine recovered as S-Ppyridylethylcysteine-Cys by compositional analysis, based on 3 mol of cysteine/mol of protein) was unaltered. Binding of Ca(I1) by S-P-pyridylethylcysteine-CaBP also was not significantly decreased (within the limits of experimental error for a single point determination), confirming the findings of Ingersoll and Wasserman (9) that sulfhydryl groups are not involved in Ca(1I) binding. These results, together with those for Pb(I1) binding to bovine CaBP (which contains no sulfhydryl groups) support the conclusion that Pb(1I) and Ca(I1) compete for the same set(s) of high affinity-binding sites.
Similarities in primary structure corresponding to the predicted Ca(I1)-binding domains among members of the troponin C superfamily of calcium-binding proteins suggested that high affinity Pb(I1)-binding may not be unique to the vitamin D-induced CaBPs. In order to test this possibility, Pb(I1) displacement of 45Ca from bovine brain calmodulin and ium-binding Proteins bovine skeletal troponin C was studied under identical conditions to those described for the vitamin D-induced CaBPs. M), were dialyzed against varying concentrations of stable Pb(II), as before. The results (Fig. 3) clearly show that Ca(I1) is displaced by Pb(I1) from both proteins and that the concentration of Pb(I1) required for 50% displacement of Ca(I1) is of the same order of magnitude (-micromolar) as observed for the vitamin D-induced CaBPs. Computation of the affinity constants was complicated, in these cases, since troponin C is reported to contain 2 sets of 2 Ca(I1)-binding sites each, with widely differing k? (17) and, whereas calmodulin is known to contain 4 Ca(I1)-binding sites in 2 sets, uncertainty exists as to the distribution of these sites and their k,'" values (16). For purposes of comparison, however, it is evident from the results in Fig. 3 that the average kaeb is greater than the auerage k? for both proteins.
Actual Pb(I1) binding (in contrast to Ca(I1) displacement) to calmodulin, troponin C, and oncomodulin, a parvalbuminlike Ca(I1)-binding protein of tumor origin (18) More detailed studies are required to completely detail Pb(I1) binding to these proteins.
The relative order of binding affinities for both the bovine (8) and chick (9) CaBPs for alkaline earth cations has been shown to be related to ionic radius, i.e. Ca2+ (0.99 A), S9' (1.13 A) and Ba2+ (1.35 A), as have the cations of the lanthanide series, in spite of the additional charge. The exceptional affinity of CaBPs for Pb(II), cannot be explained solely on the basis of ionic radius, since Pb(I1) (1.21 A) would be predicted to fall between Sr(I1) and Ba(I1) in the affinity series. Pb(II), however, can form polarized or partially covalent bonds with some ions, a property not shared by the alkaline earth cations. It is predicted, on the basis of size therefore, that the Pb(I1) ion would "fit" the binding sites normally occupied by Ca(II), but rather than being coordinated by all the carboxylate ligands in the usual fashion (Le. for Ca(II), Pb(I1) may form one or more covalent bonds with these (or other) ligands). The result would be a stabilized structure reflected by a relatively high ka value. The exact Both proteins, pre-equilibrated with 45Ca (CaF = 1.0 X 0 Calmodulin The studies reported here conclusively demonstrate specific high affinity binding to several calcium-binding proteins and suggest it to be a general property of those proteins which constitute the troponin C superfamily of "E-F hand" type calcium-binding proteins. The physiological significance of the observations, especially related to lead toxicity, remain to be studied. However, these Pb(I1)-protein interactions may well be of importance considering the broad role of free Ca(I1) ion as an important second messenger and the association of this class of proteins with a variety of Ca(I1)-mediated cellular processes.