The Binding Properties of Anti-Phosphorylcholine Mouse Myeloma Proteins as Measured by Protein Fluorescence

Ligand-induced enhancement of tryptophanyl fluorescence was evaluated in a series of mouse myeloma proteins with phosphorylcholine-binding activity. Phosphorylcholine binding by the protein TEPC 15 and its Fab’ fragment was studied. The enhancement in the Fab’ preparation is produced by an increase in tryptophanyl quantum yield, which is sensitive to quenching in alkali but not in acid. The Fab’ preparation binds 0.9 mole of phosphorylcholine with an association constant of 5.5 X 10” M-‘. The enthalpy of binding is 11 Cal per mole and the entropy of binding is 11 cal per deg mole, indicating that association results from the enthalpy change. The binding affinity is insensitive to pH between pH 6 and 9, but is markedly decreased in acid.


SUMMARY
Ligand-induced enhancement of tryptophanyl fluorescence was evaluated in a series of mouse myeloma proteins with phosphorylcholine-binding activity. Phosphorylcholine binding by the protein TEPC 15 and its Fab' fragment was studied. The enhancement in the Fab' preparation is produced by an increase in tryptophanyl quantum yield, which is sensitive to quenching in alkali but not in acid.
The Fab' preparation binds 0.9 mole of phosphorylcholine with an association constant of 5.5 X 10" M-'. The enthalpy of binding is -11 Cal per mole and the entropy of binding is -11 cal per deg mole, indicating that association results from the enthalpy change.
The binding affinity is insensitive to pH between pH 6 and 9, but is markedly decreased in acid.
The specific binding of small molecules by purified homog:eneous antibody preparations (I, 2) and by certain human (3) and mouse (4-7) mycloma proteins proTides a model for antibody binding to antigenic determinants.
The nature of the binding affinity in these systems has generally been determined by equilibrium dialysis (4,5,7), and more conveniently by the l&and-induced quenching of intrinsic tryptophanyl fluorescence (4,5) in the case of chromophoric ligands.
In this paper we report the ligand-induced enhancement of tryptophanyl fluorescencc in several mouse myeloma proteins which bind phosphorylcholine.
The increase in tryptophanyl emission may be employed to determine the binding stoichiometry, affinity, and homogeneity of thcsc and other immunoglobulin systems behaving in this manner.
This fluorescence change does not depend on energy transfer from fluorescent donors in the antibody to absorbing acceptor groups in the ligand.
The method is applied to a study of the binding characteristics of TEPC 15 and its monovalent Fab' fragment. MATERIALS AXD METHODS Xouse IgA mycloma proteins MOPC 167, XIOPC 511, McPC 603, TEPC 15, and IIOPC 8, wit,h activity toward phosphorylcholine, have been described by Potter and Leon (8) and Leon and Young (6). The mouse IgG mycloma protein ilLl'C 43 with activity toward phosphorylcholine has been isolated by Drs. Potter and Rudikoff, National Inst.itutes of Health. All the myeloma proteins and the TEPC 15 Fab' were prepared and kindly provided by Dr. Rudikoff.
The possibility of Ca '+ effects on the studies reported here was eliminated by the inclusion of controls containing 0.01 M EDTA or 0.005 M CaC12. Disodium phosphoryllmethy~-I%choline (New England Nuclear Corp.) also produced identical results.
All chemicals wcrc reagent grade. Pur$cafion 0s Proteins-Mouse Igh myeloma proteins with activity toward phosphorglcholine were purified by Dr. Rudikoff as described by Chesebro and Mctzger (9). Serum or ascitcs fluid from tumor-bearing animals was reduced with 0.005 31 dithiothreitol and alkylated with 0.01 RI iodoacetamide (recrystallized two times).
The material was then passed through Sepharose-phosphorylcholine affinity columns equilibrated with 0.20 M borate buffer-O.16 M XaCl, pH 8.0. Columns were washed with t.his buffer until the absorbance at 280 nm of the effluent was less than 0.05, and protein was then specifically eluted with 0.002 M phosphorylcholine.
Antibody-containing fractions were then dialyzed against the same buffer for removal of phosphorylcholine, concentrated, and frozen. Fab' Fragments-Pepsin Fab' fragments were prepared as previously described (10). Purified protein was reduced again in order to complctc reduction of the interheavy chain disulfide bonds with 0.01 M dithiothreitol, alkylated with 0.02 M iodoacetamide, and dialyzed against 0.1 M acetate buffer, pII 4.5. The protein was then digested with pepsin (l:lOO, enzyme to protein) for 4 hours.
Fab' fragments were separated by passing the digestion mixture through a Sephades G-100 column in the borate buffer followed by rechromatog:raphy on a Sephadex G-200 column.
Protein Concentralion-The concentrations of the TEPC 15 and TEPC 15 Fab' solutions were dctcrmined by absorbance at 280 am at neutral PI-I. The estinction coefficients were determined by differential refractometry (C. N. Wood Manufacturing Co.) of the protein solutions in water at 546 nm at 25" with the use of a refractive indes increment of 1.89 X 10e4 ml per rng (11 Ideterminations of nitrogen and agreed within 17; w&h the concentration determined on a weight basis.

Tryptophanyl and Tyrosyl
Contents-The tryptophanyl and tyrosyl contents of the TEPC 15 and TEPC 15 Fab' proteins were determined spectrophotomctrically in 6 M guanidine hydrochloride from absorption measurements at pH 6. 5 and 12.5 (12). The contents are reported in Table 1 Fluorescence Spectra-Emission and excitation spectra were measured with the Turner model 210 spectrofluorometer which gives corrected spectra in quanta per unit band width.
Excitation was at 280 nm and at 295 nm. The absorbances of the protein solutions were less than 0.05 at the excitation wave length.
The quantum yields of tryptophanyl emission were calculated by comparing the absorbances, A, at the exciting wave length and the areas of the emission spectra of the protein and a standard of known quantum yield. (15). At the higher protein concentration, small corrections for changes in midcell absorbance at the exciting wave length (295 nm), employing the antilogarithm of one-half of the absorbance, were included (16). The moles of ligand bound per mole of protein, ;, was calculated from the values of added ligand per mole of protein (CJP) which give the same value of AF/AF,n,, a t protein concentrations P, and Pb, using the relation of Halfman and Nishida (15)

x 10M5 M and FM, essentially following the method of Halfman and Nishida
The standard was a twice-recrystallized preparation of N-acetyl tryptophanamide.
A quantum yield of 0.13 has been reported for this compound in aqueous solution at 25" (13). The absorption of the tryptophanyl residues of the proteins at 280 nm was calculated from the protein absorbance, the tryptophanyl and tyrosyl contents indicated in Table I, and the molar extinction coefficients of the tryptophanyl and tyrosyl residues of 5600 and 1200 (14), respectively, at 280 nm.
For all of the other studies with this fragment, ; for phosphorylcholine binding by TEPC 15 Fab' obtained from this graph was used. TEPC 15 was considered to have 1.9 sites as measured by equilibrium dialysis (9).

Determination of Binding by Fluorescence
Enhancement-The enhancement of tryptophanyl emission provides a rapid method of determining the binding of phosphorylcholine by TEPC 15 and its Fab' subunit.
The protein solution (1.5 ml), wit.h concentration less than 1.5 p&f, was excited at 295 nm (band width, 4 nm). The emission at 330 nm (band width, 10 nm) was followed in a Perk&Elmer model MPF3 spectroiiuorometer equipped with a temperature-controllctl cell compartment. Dust was removed from the solutions by centrifugation.
The ligand was added in small volumes t.o the fluorometric cell with magnetic stirring, employing an Agla syringe with a small Teflon nozzle. The total volume of phosphorylcholine added in the entire titration was about, 100 ~1. Fluorescence standards were included to monitor any drift of the instrument.
The total enhancement in emission was determined by adding small volumes of 0.01 M or several crystals of phosphorylcholine until no further increase in emission occurred.
After corrcct'ions for dilution, the fraction of binding sites occupied was determined from the ratio of fluorescence increase to maximum increase (saturation).
The concentration of free l&and was calculated by subtracting the small amount of bound ligand from the concentration of added ligand. The emission maxima and the percentage increase in tryptophanyl fluorescence with excitation at 280 nm are given in Table II. Ligand-induced enhancement of tryptophanyl fluorescence, associated with small blue shifts of emission maxima of about 1 am, was present for all of the proteins with phosphorylcholine-binding activity.
The percentage increases in emission ranged from 10 to 25%. The MOPC 315 and MOPC 173 controls showed no change in tryptophanyl emission.
The stoichiometry of binding for TEPC 15 Fab' was determined from the nuoresccnce data at protein conccnt.rations of Emission Spectra-In order to investigate further the nature of the ligand-induced enhancement of tryptophanyl fluorescence, the binding of phosphorylcholinc by TEPC 15 Fab' (henceforth referred to as Fab'), where the number of tryptophanyl residues per binding site is reduced from 12 to 7 or 8, was studied. The emission spectra in the absence and presence of excess phosphorylcholine are depicted in Fig. 1 nm, where tyrosyl residues do not absorb significantly, the fluorescence comes only from tryptophanyl residues.
In the presence of excess phosphorylcholine, tryptophanyl emission increased 28% t'o give a quantum yield of 0.09, and the emission maximum was 2 nm blue-shifted to 329 nm.
When Fab' with and without phosphorylcholine is excited at 280 run, where tyrosyl residues absorb about 4O"/c of the exciting radiation, the wave length dependence of the emission spectra corresponds to that obtained with excitation at 295 nm. Therefore, tyrosyl emission is not observed and is quenched by other groups in the protein.
In addition, the wave length dependence of t,he excitation spect'ra of t.he protein and complex is virtually identical and corresponds to the absorption spectrum of tryptophan, indicating that there is no significant energy transfer from tyrosyl to tryptophnnyl residues in the univalent antibody or in the complex.
pE-i Dependence-The effect of high and low pH on the tryptophanyl rmission of Fab' in the absence and presence of excess phosphorylcholine is illustrated in Fig. 2. Solutions were titrated from neutral pH by adding either acid or alkali.
The fluorescence of the fragment alone is quenched in acid by 40% aud in alkali by 65c0 with midpoints near pH 4.4 and 9.9, respectively.
There is no change in the wave length of the emission maximum from pH 3.0 to 11.0, which suggests that no major structural change occurs in this pH range. Above pH 11, timedependent changes in fluorescence indicate that the protein is no louger stable. The alkali quenching transition is seen with most nat,ive proteins and usually represents the quenching of tryptophanyl fluorescence by ionized tyrosyl residues through radiationless energy transfer.
The rather strong quenching observed in acid is presumably due to quenching by neighboring protonated carboxyl groups, since antibodies appear to be relatively stable in dilute acid solutions (17). The fluorescence curve was largely reversible when the pH was increased from 3 to 7.
The fluorescence curve of Fab' with phosphorylcholine closely resembles that of the Fab' alone with the same two transitions and midpoints.
However, the acid titrations of the protein and complex reveal a constant difference in emission intensity, whereas the alkaline titrations show a continuously decreasing difference which disappears above pH 11. This indicates that the ligand is still bound at the lowest pH of the fluorescence curve and until pH 11. The binding in acid was demonstrated directly by adding phosphorylcholine to a solution of Fab' at pH 3.0 (Fig. 2). The fluorescence increased as expected, but much larger amounts of ligand were necessary to produce the same increase as was found at neutral pH. The data gave a linear Scatchard plot with a binding constant of 6.0 X 10e3 n1-l (see below). The result. for Fab' was 5.5 x lo5 &r-i, in accord with that determined by equilibrium dialysis (7). The binding constant for TEPC 15 was identical within espcrimental errors with that found for its Fab' fragment. pH Dependence-The pH dependence of phosphorylcholine binding to Fab' from pH 5 to 9 in 0.01 M acetate-O.01 M Tris-0.15 M NaCl at 25" is depicted in Fig. 4. The binding constant is relatively insensitive t.o pH between pH 6 and 9, but it is reduced by more than 50% at pH 5.0. ht pH 3.0 in 0.01 M citrate, 0.15 RI KaCl, the binding affinity is reduced about lOOfold from that at neutral pII, yielding a value of 6.0 X lo3 M-'.
Salt Concentration--The dependence of binding on NaCl and NaI concentration in 0.01 M Tris, pH 8.0, was also investigated. The Scatchard plots are shown in Fig. 5 The results (Fig. 6) show that the binding constant increases about 4-fold as the temperature is lowered from 40 to 15". h plot of log K versus l/T is linear (Fig. 6. inset). Binding-The ligand-induced enhancement of tryptophanyl fluorescence in the mouse myeloma proteins with phosphorylcholine activity provides a convenient, rapid method for studies of ligand binding.
The applicability of this method was demonstrated by measuring the binding properties of TEPC 15 Fab' under a variety of conditions. The stoichiometry of binding was determined from the fluorescence data at two widely different protein concentrations and was found to be 0.9 mole of ligand per mole of protein (Fig. 3). The fluorescence response was proportional to the extent of ligand binding. The result&z Scatchard plots under varying conditions were linear, demonstrating homogeneity of binding sites. The Scatchard plot for phosphorylcholine binding by TEPC 15 was likewise linear, with virtually the same binding constant as for its Fab' fragment. As in other hapten-binding studies (19), this indicates the absence of interactions between the two sites and rules out any large ligand-induced structural change in the Fc portion of TEPC 15.
Phosphorylcholine binding to Fab' was independent of pH between 6.0 and 9.0 (Fig. 4). It is unlikely that electrostatic interactions between the antibody and ligand contribute significantly to the free energy of binding, since large changes in ionic strength had very little effect on affinity at pH 8.0 (Fig. 5). However, the binding affinity decreased by a factor of 2 at pH 5.0 and by 100 at pH 3.0. The binding of protons by the carboxylate groups titrated in this region (17) increases the positive charge on the protein.
The greater net charge at acid pH values could lead to minor displacements of neighboring groups and impairment of some of the loci of the binding site. With the loss of binding loci, the affinity would decrease. It is also possible that protonation of the secondary acid of phosphorylcholine (pK of the ligand analogue, phosphorylethanolamine, is 5.6 at this ionic strength (20)) ma y 1' e iminate a functional group on the ligand needed for binding.
The enthalpy and entropy of the binding reaction were obtained from the temperature dependence of the binding constant (Fig. 6). The enthalpy was -11 Cal per mole and the entropy was -11 cal per deg mole. These values are similar to those reported for rabbit antibodies to e-dinitrophenyl-L-lysine and 2,4-dinitroaniline (21). At 25", the enthalpy (-11 Cal per mole) favors binding while the entropy effects decrease the negative free energy by 3 Cal per mole, resulting in a AF of -8 Cal per mole. Since phosphorylcholine has little or no hydrophobic character to account for the affinity, it is possible that hydrogen bonding is responsible for its high enthalpy of binding.
The phosphate group offers a very good site as a hydrogen bond acceptor. Fluorescence Enhancement--Since the tryptophanyl quantum yields of TEPC 15 (0.06) and TEPC 15 Fab' (0.07) are approximately equal, the increased fluorescence enhancement with excitation at 295 nm from lS70 for TEPC 15 to 28% for the Fab' is fully accounted for if the fluorescence enhancement is confined to the Fab portions of the TEPC 15 protein, and it is independent of the Fc portion of the molecule. The small blue shift of the emission maximum (2 nm) for Fab' with phosphorylcholine binding suggests that the microenvironment of the affected tryptophanyl residue(s) becomes slightly more nonpolar (22). Although this may represent decreased exposure to solvent by direct ligand shielding, other changes in aromatic chromophoric properties of the antibody leave open the possibility of a ligand-induced change in protein conformation. Further studies on this problem will be presented elsewhere.' The acid quenching curves of Fab' Huorcscence ( Fig. 2) indicate that the groups (presumably carboxyl) responsible for the 1 R. Pollet, S. Rudikoff,M. Potter,and H. Edelhoch,unpublished data. 5447 quenching have no effect on the emission behavior of the tryptophanyl residue(s) whose quantum yield(s) increase on ligand binding.
Contrariwise, the alkali quenching of tryptophanyl emission by energy transfer to the numerous tyrosyl phenolate ions is a much longer range effect than that due to carboxyl quenching.
Consequently, the difference curve in alkali becomes increasingly smaller as tryptophanyl intensity decreases, and the percentage increase with binding does not increase as in acid but remains constant.
Ligand-induced enhancement of intrinsic tryptophanyl fluorescence is present in all of the mouse myeloma proteins examined which have activity toward phosphorylcholine (Table II). Although this finding might suggest a common structural relationship among these proteins which is independent of their individual immunoglobulin type, specificity spectrum (6), idiotype (23)) and binding affinity (7), the origin of the fluorescence changes may well be different in each case since the magnitude of the observed enhancement varies from 10 to 25%.
Recently, ligand-induced enhancement of tryptophanyl fluorescence in two myeloma proteins possessin g anti-&1,6-galactan activity has also been found (24). It is unclear at present whether this finding reflects the presence of tryptophanyl residue(s) in the active sites of these molecules, or ligand-induced modification of protein conformation, or both effects.