March 20, 1961
INTERNAL RIGIDITYOF THYROGLOBULIN
alkalinized urea and guanidine solutions both the intrinsic viscosity and sedimentation coefficient were found to decrease significantly from their values in neutral solution. The latter parameters will usually move in opposite directions if configurational modifications occur without mass changes. When both constants decline simultaneously a reduction in particle size can normally be assumed. It seems evident consequently t h a t a further decrease in the average molecular weight of thyroglobulin subunits occurs in urea and guanidine solution when the pH is raised t o -12. Comments Dissociation of thyroglobulin into S-12 subunits either by raising the pH ( 1KNOa 7.0 Sone 55.0 f 10 TG-CN-IV 32.5 i 5 4.1 ,01 K S 0 3 7.0 SaC1*S04 0.001 TG-CN-I11 1.03" 4.1 . 10 KN03 7.0 XaCl2SO4 ,001 41.0 =t6 TG-CN-I11 0.8id 6.1 ,1Kxo3 12.2 Xone 1 1 . 1 =t 0 . 5 TG-CN-I1 (0. 83)'(.55)e . 0 1 K;\'Oa 7.0 ATaCI2SOc ,0029 7 . 2 =t 4 6.1 TG-CN-I1 2.5 , 0 1 KNOB 7.0 TDAC .04 3.3 f 0.2 TG-CN-111 .12" 1.3 ,001 KNOs 7.0 SaC12SOa ,018 3.3 f 0 . 2 TG-CN-IL' .12" 2.1 .01 KNOa 7.0 NaCl&304 .035 3 . 0 =t0 . 2 TG-CN-111 .11" 2.5 .01 KNO3 7.5 Urea 8.0 2.4 f 0.2 TG-CN-111 (. 16)' 2.8 . 0 1 KXOa 7.0 Urea 8.0 2.4 f 0.2 TG-CN-IV ( . 16)' 2.5 . 0 1 KNOs 11.7 Urea 8.0 1 . 4 zt 0 . 2 TG-CN-I11 1.6 HzO 6.9 GHCL 5.0 1 . 3 zt 0 . 3 TG-CY-V 4 Computed assuming M = 3.3 X l o 5 (ref. 1, 2, 4, 15); assuming M = 1.6 X 1oj; 6 assuming AP = 3 8 X l o 5 (from fractional S-12 ( M = 3.3 X l o b ) and S-19 (iM = 6.7 X lo5) content): assuming hl = 5.5 X 105 (from fractional 5-12 and S-19 content); e assuming J l = 2.4 X 105 (experimental value)4 may be too high.
observed. KO explanation is available a t present for this anomalous effect. It is possible t h a t it is an artifact arising from a slight curvature of the curves of R zwszts T / ? a t lower T/q values.
molecular entity in view of the heterodispersity revealed by boundary spreading and asymmetry of its sedimentation pattern.I5 The latter phase of molecular fragmentation is Results accompanied by moderate changes in internal A. The Effect of Individual Agents. Native structure, as reflected by alterations in solubility, Thyroglobulin.-The relaxation time of native viscosity, ultraviolet spectrum and optical rotation. thyroglobulin (mol. wt. 6.7 X lo5) was so large These latter phenomena are largely reversible that only a slight temperature dependence of its when the p H is reduced. polarization was observed for fluorescent conjuTitration of the thyroglobulin conjugate to gates of this excited lifetime; about 57, for a 20' alkaline pH's resulted in a gradual drop in polarichange in temperature was found. Thus only a zation, beginning a t about p H 10, as Fig. 1 shows. quite approximate value of is computable. The data were obtained immediately after the This is cited in Table 111. IVithin experimental adjustment of the pH to the value cited. If the uncertainty, it is in the range to be expected for a conjugate was back titrated from pH 12.3, adefinite rigid particle of this size. hysteresis appeared, reflecting the incomplete The Effect of Thermal Treatment.-Exposure of (or time-dependent) reversibility of the process. thyroglobulin (0.01 M KNOs, pH 9.8) to a thermal However, the recovery of polarizations close to the cycle ( 5 minutes a t 70') results in only a marginal initial values indicates that the rigidity of the (2y0)drop in polarization a t 25'. native molecule was largely regained. Thermally denatured solutions contain two sediIf the measurements of polarization are extended menting components : approximately 70% S-12 over a range of temperatures (Fig 2 ) a t PH 12.2, and 30y0 S-17,l4 The observed decrease in P the mean relaxation time may be computed (Table is too small to account for the formation of 707, 111). The value computed for Ph'" under these of a subunit of half the molecular weight of native conditions is 1.1 X The ratio of this quantity thyroglobulin. If, however, as is proposed else- to 37I'lKT is about 0 . S 8 (or 0 33, depending upon where,4 the S-17 component is a dimer of S-12 the value chosen for AI). Inasmuch as the probwith greater asymmetry than native thyroglobulin able deviation of the shape of the molecular domain (S-19) then the polarization of this molecule should from spherical symmetry, as well as the occurrence be even greater than that for S-19 and hence could of appreciable hydration would tend to increase cancel most of the decrease expected in P due to the this value above unity, it is highly likely t h a t this S-12 molecule. The small change in P should also value reflects the introduction of internal degrees rule out the possibility that S-17 is an unfolded of rotational freedom. In other words, some of the form of S-19. I t is clear t h a t thermal denaturation rigidity characteristic of the intact molecule has under these conditions has not resulted in any been lost. However, the effect is slight in comimportant loss in internal rigidity. parison with t h a t accompanying the action of The Effect of Alkaline pH.-The complex detergents or of urea, as will be discussed subsesequence of events occurring a t alkaline PH has quently. The above observations are in harmony been the subject of other papers in this series. The with the fact that, a t pH's about 12, the intrinsic earliest observable effect is a dissociation of native viscosity of thyroglobulin is ionic strength-dethyroglobulin into a split product (S-12) and a pendent. smaller amount of a second component (S-17). The Effect of Detergents.-The addition of Above pH -1 1.4 further breakdown occurs, lead- either sodium dodecyl (CU) or decyl (GO) sulfate ing to a slower sedimenting component, L e . s-9. to a thqToglobulin conjugate a t neutral PH and This latter component is probably not a discrete 26' resulted in a rapid drop in polarization (Fig. 3 ) . (14) H. Metzger and H. Edelhoch, J. Am. Chem. SOL.,85, 1428 (1961).
(15) Unpublished observations of H Metzger, R. E. Lippoldt and H. Bdelhoch
INTERNAL RIGIDITYOF THYROGLOBULIN
March 20, 1961
1439
0.205 0.50
I
I
1
I
I
0.01
0.02 0.03 0.04 Molarity detergent. Fig. 3.--Variation of relative polarization with detergent concentration for TG-CN-I1 (6.1 g./l.) in 0.01 M KNOI, pH 7.1,25.l0: 0, Na decyl sulfate and 0 , Na dodecyl sulfate. The vertical dashed lines indicate the reversal occurring upon thirtyfold dilution with 0.01 ;If KNOa.
6 0.195
0.20
0.18.5 10
11
12
13
6
0.15
PH. Fig. 1.-Variation of polarization (relative t o pH 7 ) as a function of pH a t 26", TG-CN-11, 5.7 g./l., in 0.1 Af KKOa: 0 , forward curve and 0 , reverse curve.
0.10 0.026 0.05 Molarity decyl sulfate. Fig. 4.-Reversibility of detergent action for TG-CN-I11 (1.1 g,/l.) in 0.01 M KKOs, pH 7.0, 26": 0, forward curve and 0 , reverse curve ( b y dilution a t constant protein concentration).
The effect was much more dramatic than t h a t observed a t alkaline pH as a comparison of Figs. 1 and 3 shows. A considerable quantitative difference occurred between decyl and dodecyl, the latter being much more effective on a mole-for-mole basis. l6 Figure 4 shows the reversibility of the process. For a particular value of the detergent concentra3.0 tion and of the detergent-protein ratio essentially 2.5 5.0 the same polarization was attained irrespective of T/T the direction from which the given molecular state Fig. 2.-Variation of 1/P 1/3 with temperature for a was approached. The reverse curve was obtained range of conditions: 0 , TG-CN-III,4.1 g./l., 0.001 M h'a do- by dilution with a non-conjugated thyroglobulin solution. Alternatively, a hundredfold dilution decylsulfate, 0.1 Af K S 0 3 ,pH 7.0; X, TG-CIi-11, 5.7 g./l., with solvent of the protein-detergent mixtures 0.1 M KNoa, pH 12.2; i3,TG-CN-Il-, 2.5 g./l., 8.0 AT urea, 0.01 M KNOS, pH 7.5; W, TG-CN-111, 2.5 g./l., resulted in the recovery of a polarization close t o 8.0 M urea, 0 0 1 M KNOB, PH 11.7; 0,TG-CIi-111, 2.1 the initial value (Fig. 3). These results do not, g./l., 0.045 Lf Na decylsulfate, 0.01 M KNOs, pH i.0; of course, necessarily indicate that the initial fine structure is completely regained upon dilution A, TG-CN-V, 0.71 g./l., 0.037 A f Na decylsulfate, 0.01 AT a
+
glycine, pH 9.7; 1.85% mercaptoethanol (v./v.).
(16) H. Edelhoch, J . P h y s . Chem., 64, 1771 (1960).
14.10 0.15
I t i c :riill t i i t i t i i i d . i t : i .
March 20. 1961
INTERNAL RIGIDITYOF THYROGLOBULIN
144 1
0.20
0.20
I
II
0.175
-i I II I
6 0.15
\
:\ I
\
0 0.15 10.0 5.0 Molarity urea. Fig. 6.--Ke>ersibility of action of urea for TG-CS-I11 (0.6 g./l.) in 0.01 M KSOa, p H 7.8, 26': 0,forward curve and 0 , reverse curve (by dilution). The data are uncorrected for the increase in viscosity with urea concentration.
The effect of 5 Ji guanidine hydrochloride was even more profound than that of 8 M urea, as Table 111shows. This is in accord with the greater effectiveness of guanidine in unfolding proteins and denaturing enzymes, on a molar b a ~ i s . ~ * , l ~ B. Effect of Reagents in Combination. Effect of Urea and Detergent.-& illustrated in Fig. 7 when thyroglobulin solutions, in varying concentrations of urea, were brought to a given level of detergent concentration, the polarization was further reduced in all instances, except at the highest urea concentration studied, i.e., 9 M . If we can assume that urea has no direct influence on the stability of hydrophobic bonds and that detergent is ineffective against hydrogen bonds, then the disruption of the hydrogen bonded groups in thyroglobulin by 9 AT urea leads t o the cleavage of most, if not all, of the hydrophobic bonds simultaneously. The converse is, however, certainly not true since the optical rotation results indicate that a considerable amount of secondary structure remains in thyroglobulin unfolded by excess detergent. It is likely, however, that detergent does succeed in destroying the hydrogen bonds formed between amino acid residues in some cases, if we can generalize from the behavior of the phenolic hydroxyl groups. The latter have been shown t o titrate reversibly in detergent solutions.4 Moreover, the ultraviolet difference spectrum elicited by detergent is almost identical to that observed in 8 -11 urea. It appears therefore that certain portions of the 0-helical backbone are independently stable and do not require any support from the tertiary structure. In 9 AI urea the polarization was almost invariant t o the addition of detergent (Fig. i ) . If there existed a class of non-covalent bonds resistant to urea, but labile to urea and detergent in (18) 3 . Schellman, R . Simpson and W. K a u z m a n n , J . A m . Chein. S o r . . 75, 5152 (1Yd3). (19) H. Edelhnch, ibid., 8 0 , 6648 (lQ.58).
0.10
3.0 5.0 7.0 9.0 Molarity urea. Fig. T.-Comparison of effects of urea alone (0) and with 0.0025 M sodium dodecyl sulfate ( 0 ) for T G - C S - I 1 (2.4 g./l.) in 0.01 ,Ilr KS03, 26". Both curves are corrected for the increase in viscosity with urea concentration, using equations 1 and 3. 1.0
combination, it would be expected that the combined action of these reagents would reduce P h to a lower value than 9 i V urea alone. Since this does not occur, it appears likely that no such class of bonds exists. Examination of the temperature dependence of polarization in 8 hf urea revealed that p h Z 5 was greatly reduced from its value for the intact molecule and was, in fact, even smaller than the limiting value found in excess detergent (Table 111). 2 0 The difference between the two values is, however, somewhat smaller than would be expected from the major differences in optical rotation in the two media. Effect of Detergent Plus Reducing Agent.As Fig. 8 shows, the polarization of the thyroglobulin conjugate was essentially unaffected by the addition of sulfhydryl reagents in the absence of detergent.21 Nor was any effect observed a t p H 7, with or without detergent. However, in the presence of detergent at slightly alkaline p H (9.6) a definite effect was observed, and the polarization fell t o a level which was dependent upon the detergent concentration. The polarizations recorded in Fig. 8 were measured immediately after the addition of the sulfhydryl reagent. The oc(20) It is worthwhile t o mention t h a t in addition t o its direct effect upon t h e molecular properties of thyroglohulin, t h e presence of urea or guanidine in high concentrations also considerahly reduces the excited lifetime (Table 11) and increases t h e solvent viscosity. Both these effects serve t o increase P a t constant temperature and hence t o counter t h e effects of t h e molecular changes discussed above. T h e d a t a of Fig 6, 7 a n d 9 are uncorrected for these factors, which partially mask t h e full extent of t h e effect. T h e cited values of p i l g i have, of course, been computed with d u e regard t o these factors. ( 2 1 ) This does n o t , of course, preclude t h e possibility t h a t an effect might be observed a t neutral p H a t sufficiently high levels of sulfhydryl reagent or for sufficiently long reaction times,
1442
Vol. sx
I, Mandelkern. J . A m . Chenz. Soc., 76, 179 L lWj3>
1701.
83
is attained in the case oi detergents. The same limiting state appears to be attained for al! three detergents tested which included both cationic and anionic detergents. I n each case the effective rotational kinetic unit approached is about 1/10 the size of the intact molecule The size of the rotational kinetic unit is essentially independent of ionic strength, in contrast to the pronounced inflation of the over-all molecular domain a t low ionic strengths, as measured b y intrinsic viscosity. Furthermore, the loosening of the structure of the submolecules in the presence of detergent appears to render them susceptible to the action of reducing agent. Rupture of the exposed S-S bridges under these circumstances results in the production of a smaller rotational kinetic unit. However, even after reduction, considerable rigidity remains. Further work will be required to determine whether all S-S bridges are ruptured and whether the observed molecular changes reflect an actual splitting of the submolecules into smaller translational kinetic units or a further loss of internal structure. The prior action of alkali does not appear t o affect markedly the susceptibility of the submolecules to detergent. I n view of the only small changes in the optical rotation produced by detergent, it appears likely that the helical substructure of the molecule remains largely intact At this stage of our knowledge the most valid picture of the limiting state of thyroglobulin in the presence of high detergent concentrations is probably t h a t of numerous short helical regions separated by amorphous regions and further constrained by S-S bridges. The limiting states attained in 8 Allurea or 3 Jf guanidine appear t o correspond t o a more profound alteration in substructure than in the detergent case, as is reflected by the smaller effective relaxation times attained under these conditions. I n view of the optical rotation results this probably reflects the destruction of the helical structure of the subunits. That all rigidity has not been entirely lost in Y 31urea or 5 -11guanidine at neutral pH is indicated by the further profound decline in relaxation time occurring a t alkaline pH. Both urea and guanidine render thyroglobulin much more sensitive to the effect of alkali than the native molecule, a property which is not shared by detergent. I n the absence of urea or guanidine the effect of alkali is much less pronounced and can be largely accounted for i r i terms of a splitting into submolecules, although some loss of internal structure does occur
