June 5, 1962
MOLECULAR PROPERTIES OF 7-GLOBULINS
2133
[CONTRIBUTION FROM THE NATIONAL INSTITUTE OF ARTHRITIS AA-D METABOLIC DISEASES,NATIONAT INSTITUTES OF HEALTH, BETHESDA,MARYLAND, AND NAVALMEDICALRESEARCHINSTITUTE,KATIONALNAVALMEDICALCENTER,BETHESDA, MARYLAND]
Structural Transitions in Antibody and Normal y-Globulins. Properties
I. Molecular
BY HAROLD EDELHOCH, R. E. LIPPOLDT AND R. F. STEINER RECEIVEDOCTOBER 27, 1961 The molecular behavior of bovine and rabbit ?-globulin has been explored by velocity sedimentation, viscosity, optical rotation, solubility, fluorescence and ultraviolet spectrophotometry. All the methods reveal a time-dependent loss in native structure whose rate becomes measurable at p H -11 and increases rapidly with pH. Only a partial recovery in molecular properties was found on neutralization. 8 -1.I urea produced considerable inflation of the molecular domain and an increase in levorotation of the y-globulins. However, further structural disorganization was produced by alkali (in 8 urea solutions). Where configurational modifications were found by the above methods, changes in fluorescence polarization were always evident. Of interest, however, is t h a t changes in polarization were detectable before any of the more classical methods were able t o record any effects. Finally evidence based on alkaline denaturation is presented which would appear t o show that a purified rabbit antibody shows the same pattern of heterogeneity as normal rabbit y-globulin.
At present i t cannot be claimed that the y- gest, the occurrence of an inflation of the molecular globulins are among the proteins which have been domain under these conditions. This in turn sugwell-characterized on the molecular level. The gests that the transition to an expandable state heterogeneous nature of this protein undoubtedly may be accompanied or preceded by more subtle imposes severe obstacles upon studies with the changes which result in a partial loss of the rigidity usual techniques. 1--4 Nevertheless the limited characteristics of the molecule. l2 Noreover, the data available permit some generalizations as to alteration in molecular-kinetic properties is paralthe behavior of y-globulins from a variety of mam- leled by changes in optical rotation which are in the direction expected if a loss in a-helical content malian sources. At low ionic strengths the y-globulins have occurred.l3>l4 limited solubility in the isoelectric region (pH It is the purpose of the present set of papers to 6.5-7.5). Examination of the material in solution examine in detail the molecular state of y-globulins a t ionic strengths of 0.01 or less has generally from two sources under a wide range of conditions. revealed the presence of extensive a s s ~ c i a t i o n . ~ .In ~ addition, particular attention has been given This has often been regarded as reflecting electro- to correlating data obtained by fluorescence static interactions between oppositely charged polarization" with other methods as explored in components of the y-globulin solutions. At a this communication. higher ionic strength (0.1) the aggregation is Methods and Materials largely s u p p r e s ~ e d . ~ ~ ~ Physical Measurements.-Ultraviolet spectra rvere While there is no evidence for any molecular transformation occurring between the pH limits measured with a Beckman DU spectrophotometer, equipped a photomultiplier attachment. Ultracentrifuge meas4 and 9, provided that the concentration of elec- with urements were made with a Spinco Model E ultracentrifuge. trolyte is high enough t o suppress association, I'iscosity determinations were made with a low shear the exposure of y-globulin to pH's outside these viscometer of long solvent time (300 seconds for water). limits appears to result in structural modifications. d Radiometer Model T T T l p H meter, equipped with Miniature Leeds and Northrup electrodes, was used t o measure These are reflected by increases in pH. The instrument was standardized with Beckman bufffrictional ratiola'lland ers of p H 4.00, 7.00 and 10.00. I n the spectrophotometric While the detailed character of the molecular titration experiments y-globulin (-O.SyG) in 0.20 M KC1 events responsible for these effects is of course un- was diluted about 5-fold into a 0.20 M KC1-0.015 M lysine solution. Approximately 2.5 ml. samples were titrated in 1 known, it is possible to make certain rather obvious cm.2 quartz cuvettes with very small amounts of either 2 M conclusions. The changes in frictional ratio a t ex- HC1 or KOH which were delivered through a polyethylene tremes of pH are certainly consistent with, and sug- catheter from a Agla precision syringe. Solutions were mixed (1) R. A . Alberty, J . A m . Chein. Soc., IO, 167.5 (1948); J . P h y s . Colloid Chem., 62, 217, 1315 (1948); 63, 114 (1849). ( 2 ) J . R . Cann, R. A Brrwn and J. G. Kirkwood, J . Bioi C h t w , 181, l i j l (1949). (8) J . W. Williams, R . L. Baldwin, W hl. s u n d e r , and P G. Squire, J . A m Chem. S o c . . 74, 1542 (195'2). ( 4 ) J . Cann, i b i d . . 76, -1213 (1953). ( 5 ) R . Lontie and P . R. lMorrison, quoted by P . Doty and J. 'r. Edsall in "Advances in Protein Chemistry," VI, 70 (1951), also unpublished d a t a of R.Lontie and P. R. Morrison. (6) G. Scatchard, A . Gee and J. Weeks, J P h y s . Chem., 6 8 , 788 (1954). (7) J. Yang and J. Foster, J . Am. Chem. Soc.. 77, 2371 (1955). ( 8 ) B. Jirgensons. Avch. Biochem. Biophys., 41, 333 (1952); 48, 1.54 (1954); 94, 59 (1961). (9) R. Phelps and J . C a m Biochinz. el B i o p h y s A c t a , 23, 149 (1Y37). ( I O ) J. Cann. J . A m Cheni. S o c . , 1 9 , 7ij1 (1957) (11) L . A . Sternberger and 11 L. Peterman, J I r n i r z i i i i i i l , 6 1 , 2Ooi (1951).
by magnetic stirring. In this way a p H curve could be obtained on a single sample. Measurements of fluorescence intensity were made with an .kminco-Bowman spectrofluorometer, modified to permit temperature control. Solubility Measurements.-Considerable caution must be exercised in the choice of buffer that is used to quench the denaturation reaction. It was initially found that partial precipitation of denatured rabbit or bovine y-globulin could be brought about by standing for several hours in 0.4 M phosphate, p H 6.8. Much more rapid and quantitative precipitation occurred if more concentrated phosphate and a protein level not lower than 0.25Yc were used. The usual quenching procedure was to mix equal volumes of globulin solution and phosphate buffer. .4 molarity of 4.5 was found to be optimal for the latter. Lower molarities (4.25 or less) ( 1 1 ) R F. Steiner and H . Edelhoch, J Arri. ( h e m Sor., 84, 2139 (lYti2). f l 3 ) J. Yang and P. Doty. i b i d . 79, 7631 (1957). (111 C. Schellman and J . A . Schellman C o i i i p t . r e i i i i I.nb C o y i s b e r g , .Sei f h i i n . , 30, I6.i ( l B X 3 ) .
2134
60 E ' 0
x
50
3 .
R v,
40
30 I
201
I
1 I1
180'
50
10
,,
*An&r&T&o-
MINUTES
I
-io.,
a
12
PH-
Fig. 1.-Influence of alkaline p H on the sedimentation coefficient of 7-globulins. 0 , bovine y-globulin in 0.2 M KCl, 0.025 M lysine, conc. = 0.507,; 0 , rabbit antibody in 0.1 M iXaC1, 0.035 JII lysine, conc. = 0.207,; 0 , rabbit yglobulin in 0.1 M iYaC1, 0.035 A f lysine, conc. = 0.507,. I
;_I
I
1
I
!
Fig. 3.-Influence of alkaline p H on the rate of increase in levorotation a t 350 mp of bovine y-globulin (0 5 % ) . Solutions containing 0 10 ill KCl, 0 10 M K2HP04 were adjusted with KOH to indicated p H ; T , 25'.
I
8M UREA
02
O I L
I I 6
P 200c
-
180
I H20
I
I 8
I
I IO
I
I
12
PH,
Fig. 2.-Influence of alkaline p H on the reduced specific viscosity of bovine y-globulin in various solvents. Protein concn. = 0 1870; T = 25 0": 0,0 02 M KC1, 0 02 Id lysine; 0 , 0.04 Af T D A C ; 0 , 8 izI urea, 0.02 M KCl, 0 02 M lysine. failed to precipitate the denatured protein maximally; higher molarities (5.0 or more) precipitated some native globulin. The precipitate was centrifuged down by 10 minutes spinning ill a Sorvall centrifuge at 2000 g. An aliquot of the supernatant was diluted with 8-10 volumes of 2 A I KOH and its absorbancy measured at 287 mp. An extinction coefficient of 13.8 was used for a lY0 solution of bovine yglobulin in the latter solvent at 287 mp. Materials.-Rabbit and bovine y-globulins were obtained from Pentex Incorporated and from Armour, respectively. Apart from a small ( < 5 7 , ) fraction of more rapidly sedimenting material, all of both preparations sedimented as a single homogeneous peak with Sz0 7 svedbergs. Rabbit anti-thyroglobulin antibody was prepared from antisera by the method of Metzger and Edelhoch.'5 Sodium dodecyl sulfate (SDS) was a purified preparation donated by E. I. du P o n t de h-emours and Company. Trimethyldodecyl ammonium chloride (TDAC) was obtained from Armour and Company. Urea was recrystallized from ethanol. Salts were reagent grade. Glass-redistilled water was used in preparing all solutions. The preparation of fluorescent conjugates of y-globulin with l-dimethylamino-naphthalene-5sulfonyl chloride ( D S S ) is described in the companion paper.'a
Experimental Results I. Influence of p H in Aqueous Media A .
Molecularkinetic Properties.-.4s seen in Fig. 1 the sedimentation coefficients of y-globulins of either species and rabbit anti(15) H. Metzger and H. Edelhoch, Nature, 193, 276 (1962).
50
MINUTES.
Fig. 4.-Influence of acid p H on rate of increase in levorotation at 350 mp of bovine ?-globulin (0.5070). Solutions containing 0.2 M KCl. 0.015 hf lysine were adjusted with HCl to the indicated p H ; T,25". body in dilute KC1 showed no important change until about p H 11. At higher pH's a decrease in SZOoccurred. T h e change was somewhat more drastic in the case of the bovine protein. There was no resolution into discrete components in either case, although boundary spreading mas enhanced. A definite species difference appeared to exist with regard to reversibility. A return to p H 9.5 after a five minute exposure at p H 12.1 in the case of the rabbit antibody resulted in the reconversion of almost half of the material into a component of sedimentation rate close to that of the original. The balance appeared as poorly resolved, rapidly sedimenting aggregates. I n the case of the bovine protein not more than about 1056 regained the sedimentation coefficient of native y-globulin. At the intermediate p H of 10.6 aggregation was somewhat less pronounced and about 807, of the rabbit and 507, of the bovine y-globulin had sedimentation properties characteristic of their native forms. Allowing the reversed solution to stand for 20 Iir. had little effect on the degree of reiTersa1. Similarly, variation of ionic strength between 0.02 and 0.20 appeared to exert no significant influence. T h e effect o f p H on the reduced specific viscosity of buvine ?-globulin in 0.02 111 KCl is reported in Fig. 2 . The viscosity did not vary between p H 6 and 10. Above p H 10 a pronounced increase was observed. The change in viscosity precedes that of sedimentation but lags behind that of the gH-profile uf fluurescence polarization reported in the accompanying paper.'* The difference between the p H profiles of viscosity and sedimentation probably arises from the greater sensitivity of the former. B. Optical Rotation.-The effect of alkaline p H on the rate of increase in levorotation of bovine y-globulin in 0.1 At KCl-0.1 X K H z P O ais illustrated ill Fig. 3. The datu reveal a number of noteworthy features. It is evident that rotatory changes commence only above p H -11 and therefore, in this instance, are less sensitive as a n indicator of nlolecular rearrangement than either viscosity or polarization of fluores-
2135
MOLECULAR PROPERTIES OF y-GLOBULIKS
June 5 , 1932
100
I
-
I pH.325 I
I
dm 5
4
i ,5CLL_~..I.-L--L-I_I_. 2 4 6
6
?H
1 0 1 2
I
I
20
‘OMINUTES.
Fig. 7 -Influence of acid pH on the rate of denaturation of bovine 7-globulin in 0 50 -14‘ KC1, 0 025 ,\flysine. Protein Fig. 5.--Influence of pH on the specific rotation of bovine conc. = 0 SOY0; T, 25.0’. y-globulin solutions (0.Z070). 0 , 0.10 M XaC1 for p H 4.0-10.4; 0.2 M KCl, 0.015 M lysine a t other pH’s (taken from Figs, 3 and 4 at 1200 minutes); e, heated t o -70” for four minutes in 0.10 S a C l (measured at 2 5 ’ ) ; 0 , 0.03 111 SDS. 1
4
I
I
I
I
I
IOOL
I
I
110
I
2
I 4
I
6
I
1 8
120
I
2
I
.4
I 6
8
PH *
Fig. 6.-Influence of alkaline pH on the rate of denaturation of bovine y-globulin in 0.20 M KC1, 0.025 M lysine. Protein concn. = 0.5070; T , 25.0”C.
Fig 8.-Influence of alkaline p H on the denaturation of y-globulin. All solutions contain 0.10 M KC1, 0.35ojO protein. The procedure was identical for each experiment (see text for details): 0 and e, rabbit antibody, obtained from single animal (AQ-25, see ref. 15). Open and filled circles were two independent experiments; A, rabbit y globulin; 0 , bovine y-globulin; T , 25’.
cence data. At higher pH’s a n initial rapid increase in specific levorotation occurs, whose magnitude is strongly dependent on p H . Subsequently, a further time-dependent increase takes place which appears to approach a limiting value which increases with increasing p H . Figure 4 reports the kinetics of the levorotatory change of bovine 7-globulin in 0.2 M ICCl a t pH 1.80 and 2.30. The limiting values of specific rotation a t these pH’s are in close agreement with those found a t 11.23 and 12.00. At p H 3.00 a rapid initial increase occurred, followed by a slow change which did not reach the value observed at p H 1.80 even after 20 hr. At p H 3.45 only a very slow rate was recorded. As with the alkaline data the earliest perceptible changes in polarization of fluorescence occur a t somewhat more neutral pH values than in the case of specific rotation. Consistent with the only partial recovery of the initial sedimentation pattern of :.-globulin, after exposure to pH 12.1 for five minutes, is the incomplete reversibility, as determined b y optical rotation. Values close to -2.55’ were observed f(>r[a1350aiter several minutes, when the p H was decreased from 12.1 to 11.0. Xt pH 10.1, [ a j a 5 0 returned to -235’ in ten minutes a i d showed no subsequent change. When the pH was taken directly to 9.2, a smaller recovery was observed, in that the rotation decreased only to -250”. Reversal of pH to 3.7 after 5 minutes a t 1.9 resulted in only a very small recovery in optical rotation. This lack oi major reversibilitp observed after exposure t o cstreines of pH is also seen after heating solutions of bovine 7-globulin to -70” for about 5 minutes. Between p H 4 and 10 the protein precipitated from solution, precluding any optical rotatory observations. Outside this range the observed values were slightly more levorotatory than those found a t 25’. The levorotation increased slightly with increasing acidity or basicity. These data are reproduced in Fig. 5 along with the rotation data observed at 25” after 20 hr
C. Denaturation.-Denaturation will be interpreted in this communication exclusively in its classical sense: as the loss in solubility t h a t occurs in y-globulin at or near its isoelectric point. T h e rate of denaturation in alkaline solution in 0.2 M KCl is shown in Fig. 6. It is abundantly clear that the data a t the various p H values cannot represent a simple process such as the single step denaturation of a homogeneous substance. A11 the kinetic curves tend to plateau at different degrees of denaturation depending criticaily on the pH. I n addition, a t p H values above 11.25, the curves do not extrapolate to 1007, undenatured protein a t zero time. I t is of interest to note that the over-all character of the solubility curve, with respect to the pH-dependence of the relative rate, is in rather close accord with the rotatory data. The changes that are evident from polarization and viscosity a t lower pH’s presumably reflect minor and reversible changes in protein conformation. T h e reversal of pH to 10.2 after exposure of bovine 7globulin to p H 12.1 for 5 minutes produced about a 20% increase in solubility which occurred within 20 minutes. The denaturation of bovine y-globulin in 0.2 M KC1 by acid is reported in Fig. 7. Again the close similarity of the solubility to the optical rotatory data is immediately apparen t. Since rate curves which covered the complete range of loss in solubility could not be obtained a t a single p H a different procedure was used to obtain the p H profile of the loss in solubility. In this method the p H was rapidly adjusted to a given value and a n aliquot of solution was removed and quenched immediately (
