The Proton Relaxation Enhancement Properties of Concanavalin

SUMMARY The measurement of the proton relaxation rate (l/Tl) enhancement (Ed*) for the concanavalinA (Con A)-Mn2f complex as a function of time following the addition of Mn2f indicates a pH-dependent rate of structural rearrangement at the bound Mn*f site. The addition of excess Ca2+ interrupts this kinetic process and results in a reduced time-independent value of Ed*. The further addition of a-methyl-D-glucoside to the Con A-Mn’+-Ca2f complex &o alters the observable EI*. Using the recently published 2 A resolution crystal structure of Con A, the results are interpreted as arising from the gradual folding of the loop of residues 12 to 22 over the initial Mn2f binding site (formed by residues 8, 10, and 24) to close the Mn z+ off from free access by water molecules.


SUMMARY
The measurement of the proton relaxation rate (l/Tl) enhancement (Ed*) for the concanavalinA (Con A)-Mn2f complex as a function of time following the addition of Mn2f indicates a pH-dependent rate of structural rearrangement at the bound Mn*f site. The addition of excess Ca2+ interrupts this kinetic process and results in a reduced timeindependent value of Ed*. The further addition of a-methyl-D-glucoside to the Con A-Mn'+-Ca2f complex &o alters the observable EI*. Using the recently published 2 A resolution crystal structure of Con A, the results are interpreted as arising from the gradual folding of the loop of residues 12 to 22 over the initial Mn2f binding site (formed by residues 8, 10, and 24) to close the Mn z+ off from free access by water molecules.

Collcanavalin
A, a protein isolated from the jack bean (1, 2), is OIE of a number of plant lectins shown to bind to and agglutinate transformed cells (3,4). The specific action of Con A' is thought to involve the binding of the lectin to a particular carbohydrate moiety exposed at the transformed cell surface (5, 6). Nontransformed cell cultures exhibit a range of abilities to bind Con A, but are not agglutinable unless previously subjected to mild proteolytic action at the cell surface (6). A more detailed understanding of the structural features of Con A, and the molecular basis of its interaction with the carbohydratecontaining receptor, would be of value in assessing proposed models and interpreting the cell surface changes accompanying transformation.
The subunit composition of purified Con A has been described by Wang et al. (7). The structure was found tn have an intact subunit molecular weight of 27,000. Sedimentat,ion data indicate that in the pH range 3.5 to 5.8 Con A exists as a dimer of molecular weight approximately 55,000. At pH 6.0 and above, higher molecular weight forms are seen (8,9). Agglutination of transformed cells by Con A can be competitively reversed by certain carbohydrates.
The saccharide binding nbilit,y, and the agglutinability of transformed cells in culture, i:: dependent upon divalent metal ions (3, 10). Kalb  but not CaZf or Mg2+. Site S2 appears to be highly selective for Ca2+, but is only available for occupation upon previous binding of Ni2+ or Mn2+ in the Sl site. As well as pointing out this unique example of "site induction," Kalb and Levitzki concluded that both Sl and S2 sites must be occupied in order to obtain significant binding of cY-methyl-D-glucopyranoside.
As part of the effort in our laboratory to develop useful NMR probes of cell surface structure, we have investigated the solvent proton relaxation enhancement (12) properties of the M&-Con A complex.
In this technique the fact is utilized that Mn*+, a paramagnetic ion, considerably shortens the spinlattice relaxation time, T1, of the protons of Hz0 molecules in the hydration sphere of the ion, and does so differentially when the ion is bound to a macromolecule or free in solution.
Since the proton spin-lattice relaxation rate, l/Tl, is usually increased in the presence of a specific Mn*+ binding macromolecule, a spinlattice relaxation enhancement parameter, el*, can be defined 02). el* represents the ratio of the paramagnetic ion contribution to the proton relaxation rate in the presence and absence of the macromolecule for a specified paramagnetic ion concentration. A consideration of the contributions to the water relaxation time enables one to write CI* in terms of three parameters Ebl, ef, and xb. Thus, where xb represents that mole fraction of the total Mn2+ that is bound to the macromolecular sit.e and ej has the value 1 assuming no differential viscosity effects between the Mn2+ hydrate in the macromolecular solution and the buffer alone. Ebl iS a number which is characteristic of the complex which the Mn2+ forms with the macromolecule, in this case Con A. In a completely analogous manner, the enhancement parameter eb2 can be defined for the spin-spin relaxation rate, 1/T,.

EXPERIMEKTAL PROCEDURE
The buffers used in this study, 0.05 M sodium acetate in 0.2 M NaCl, were prepared in glass distilled Hz0 using BDH Analar grade NaCI, sodium acetate, and acetic acid. Ca2+ was provided in the form of CaClz. 2 Hz0 (BDH Analar) and Mn* as MnClz. 4 I-I20 (Fisher Certified), and ar-methyl-n( +)-glucoside was obtained from BDH Biochemical.
The M& stock concentration was 1.25 mM in distilled H20. Con A was obtained from Sigma Chemical Company as a lyophilized powder (Grade IV, salt-free, essentially free of carbohydrates, Lot lOlC-5390). Sodium dodecyl sulfate polyacrylamide gel electrophoresis of the Sigma product showed the same four bands, with similar relative intensities, report'ed by Wang et al. (7). Con A concentrations were determined using the optical density at 280 nm and the extinction coefficient, E'% 1 cm, of 11.4 reported by Agrawal and Goldstein (13). Stock solutions of Con A prepared with no prior treatment were found to have longer H20 solvent T1 values (-800 ms for 16 mg per ml) than solutions of similar Con A concentration subjected to the demetallization procedure of Kalb and Levitzki (11). Comparison of the ESR spectrum of the Mn*+ bound to Con A in the stock solution (14)   that fewer than 1 in 18 Mn 2+ binding sites were occupied by Mn2f in the stock solution.
Consequently, for the results reported, Con A was used as obtained from Sigma. All NMR measurements reported were made at 32 MHz on a Bruker variable frequency pulsed spectrometer  at the ambient probe temperature of 24". Ti values were determined from a semilogarithmic plot of the signal amplitude following a 1804T-90" pulse series as a function of the pulse separation 7. Samples were esamined in glass tubes. (7.5 mm diameter, Wilmad 513B-1PP). The samples used were 60 ~1 of stock Con A solution with 5+1 additions of the appropriate metal ion or sugar stock. The length of time required to obtain the T1 value for each point in the kinetic experiments was 1 to 2 mm. ESR spectra were obtained at room temperature (24") on a Varian E-6 spectrometer operating at 9.1 GHz (X-band).
Each sample (~50 ~1) was taken up in a glass disposable microsampling pipette (Corning Glass Works), which was then inserted into the microwave cavity in a reproducible position.

RESULTS
When Mn2+ is added to a buffered stock solution of Con A the relaxation time T1 is found to change appreciably with time. The enhancements obtained at three different pH values are given in Fig. 1 as a function of time following the MnZf addition. A pH dependence of the rate of change of ei* is clearly evident. Mn* ESR 'spectra were obtained for a Con A solution of 10.9 mg per ml (404 PM Mn2+ sites assuming one site per 27,000 mol wt) at a MnCl2 concentration of 96 PM from 15 to 80 min after the addition of the Mn2+.
No free Mn2+ signal was observed. The estimated lower limit of detection was 8 PM.
In view of this result, and the fact that all Con A concentrations used in the NMR experiments were considerably greater than 11.0 mg per ml and that Mn2+ concentrations were less than 100 C(M, * values reported can be regarded directly as ebl values (i.e. Tik, = 1).
The result of plotting log,, [cl* (t) -EI* (m)] versus time for the data presented in Fig. 1 reveals the possibility of two distinct first order rate processes. The estimated infinite time values for Q* used in the calculation are presented in Table I. The initial faster rate observed for each pH is at the limit of resolution for this observation technique, but the second rate process can be clearly followed and a rate constant obtained. The values calculated are 5.10 X 10m3 mine1 and 5.05 X 10m2 mine1 for pH values of 5.90 and 5.67, respectively. Kinetic data at pH 5.15 for Con A concentrations of 17.4 mg per ml and 13.0 mg per ml indicate, within experimental error, a lack of Con A concentration dependence for the transition. Addition of Ca2+ at a concentration (2.3 mM) greatly in excess of the sites' concentration at some point after the Mn2+ addition to the the Con A solution results in the behavior shown in Fig.  2. Also depicted in Fig. 2 is the effect of adding cr-methyl-nglucoside to the Con A in the presence of Mn2f and Ca2+. All values indicated in the figure have been corrected for the sample dilution and are normalized to the Con A concentrations indicated in the legend.
Further additions of Ca2+ or &methyl-nglucoside in their respective cases were found not to change the or* within experimental error, and consequently the initial concentrations used were considered saturating. A summary of the or* values obtained in the presence of Ca2+ and ol-methgl-nglucoside appears in Table I. The addition of saturating Ca2+ at any time after the Mn2+ addition interrupts the kinetic process occurring with the Con A-Mn2f complex alone, and reduces Ed* rapidly (within 5 mm) to a value which is time-independent.
ESR results indicate that upon the addition of Cazf in this situation there is no significant change in the fraction of Mn2+ bound to Con A. At each pH studied the addition of or-methyl-nglucoside to the Con A in the presence of ;\~I@ and Ca*+ further reduced the observed EI* value (Table I), and the resultant value was also time-independent.

DISCUSSION
The consistent decrease in the ~1 value observed upon addition of cY-methyl-n-glucoside to the Con AMn2+-Ca2+ complex suggests that the relaxation enhancement parameter may be useful in probing the interaction of Con A with cell surface components.
Recent x-ray diffraction data (15, 16) indicate that the saccharide binding site is at least 20 A from the bound Mn2+.
This suggests that the reduction of ebl caused by the binding of a-methyl-n-glucoside is not the result of a direct interference with the bound Mn2+ hydration sphere, but rather a more distant structural perturbation.
The detection and characterization of such rearrangements with ligands analogous to the cell surface receptors should be important in understanding the structural basis of the lectin-cell surface interaction. Equation 1 requires that the time dependence of er* derive from variations in either xb, ebl, or both. The absence of a free Mn2f ESR signal under conditions where ei* is decreasing indicates that the time dependence is not due to variation in xb but rather must arise from a change in Ebl. Of the parameters which determine Ebl (la), &*, TIMb, and 7Mb are the only ones likely to be changing with time. &*, rMb, and TM are the number, lifetime, and T,, respectively, of the water molecules in the first hydration sphere of the bound Mn2+. Regardless of which is changing, the protein must be undergoing a conformational change since all three parameters are determined by the precise nature of the water-Mn2f-protein interaction. At this time, without the knowledge of the radio frequency dependence of the time-dependent tl*, one cannot assess the possible contribution of a structural change influencing the magnitude of r&0,. However, the alternative of a change in &* contributing significantly to the decrease in EZ* is particularly attractive in the light of the 2 A crystal structure of Con A recently described by Edelman et al (15).
Examination of the stereo diagrams for Con A (15) suggests that the NH&erminal 10 or 12 residues of the Con A backbone are buried within the structure, whereas residues 12 to 22 form a loop on the surface which almost closes the Mnz+ site off from the bulk solvent.
It seems most likely, therefore, in view of our results, that Mnzf initially binds at a site on the surface of the protein involving only the side chains of residues 8, 10, and 24 and perhaps the water which is hydrogen-bonded to the carbonyl of residue 32, while the loop of residues 12 to 22 remains out in solution.
Thus, the bound Mn2+ ion would have at least 3355 two coordinated waters in rapid exchange with bulk solvent water.
Such a situation is consistent with the initial enhancement values of ebl equal to 11 which we observed.
The slow reduction in Ebl with time could then correspond to the reduction of &* in an increasing proportion of molecules as the loop of residues 12 to 22 folds over the MI?+.
In the folded structure the side chain of residue 19 would have displaced one of the coordinated water molecules, and the water hydrogen bonded to the carbonyl of residue 32 would be trapped inside the protein.
The sixth coordination position would be occupied by the water molecule at the bottom of the channel described by Edelman et al (15)) and under this circumstance may well have a ?-Mb significantly lengthened from t,he initial bound Mn2+ state. Under such conditions one would predict that the enhancement would be very small as observed.
The pH dependence of the rate of change in er* (Fig. 2) (i.e. the rate of loop folding over the Mn2+) suggests the involvement of a residue with a pK in the range of 5.8. The protonation of this residue (possibly the Mn2+ ligand His 24) accelerates the rate of folding as does the addition of Ca2+. Presumably the presence of Ca*+ further favors the completely folded structure in that it derives ligands from both the bulk of the structure (residue 10) and the mobile loop region of residues 12 to 22 (namely residues 12, 14, and 19) (15). A study of the radio frequency dependence of the time-dependent relaxation times should yield the changes, with time, in the number of coordination positions open to water, and also shed some light on the possible contribution of a change in i-Mb resulting from the structural transition.
Such a study is in progress.