Nov., 1960
1771
KOTES
&
s ___-
kiPSOIsbs
(3’)
which is kinetically indistinguishable from eq. 3. An analysis similar to that mentioned in the preceding paragraph leads to an estimate for the ratio Ica’/kt’ in the range 10-“10-*. The observed N14/N15 isotope effect is in the “normal” direction (i.e., N14 enrichment) and shows no systematic dependence on pressure (Table I). The scatter in the isotope data is worse than the precision of any individual analysis (f0.001) and is indicative of some complexity in the reaction not yet understood. Explicit expressions for the observed isotope effect (in terms of kinetic isotope effects for the various elementary steps in the mechanisms) are obtained readily. They are not reported here since data required for them evaluation are not available. tatively, both mechanisms yield isotope effects in the normal” direction. However, quantitative interpretation of the observed isotope effect is not possible a t the present time
Acknowledgment.-The authors thank Mr. J. N. Doshi for his assistance with the mass spectrometric analyses.
I
,
1
n
0.I
“0
0.2
0.3
ML.SDS. Fig. 1.-The influence of pH and deter ent concentration on the acid binding properties of thyro8obulin in 0.01 Jf KNOI. Solutions contain 3.0 ml. of 0.75% protein. Concentration of HC1 was 0.097 M and SDS was 0.10 M . Inset shows effect of pH on the maximum acid binding values; T = 28.1’.
THE PROPERTIES OF THYROGLOBULIN. 111. THE TITRATION OF THYROGLOBULIN I N SODIUM DODECYL SULFATE BY H. EDELHOCH CZinLcal Endocrlnology Branch, Xational Inatitute of Arthritis and Metobolic Diseases, Nafional Instztutes of Health, Bethesda, M d . Recezred May 23,1960
Steinhardt,’ et al., have shown that the acid segments of the titration curves of wool protein are markedly influenced by the nature of the conjugate base of the acid titrant. Putnam and Neurath? subsequently demonstrated that the binding of Duponal (a mixture of long chain alkyl sulfates) was observable by a pH shift in the protein solution. Scatchard and Black3 have formulated a relation between pH changes and the binding of anions and in this way measured the affinity of a large number of anions to human serum albumin. Sodium dodecyl sulfate (SDS) increases the pH of thyroglobulin solutions. The influence of SDS on the acid segment of the titration curve of thyroglobulin is evaluated in this report. A comparison is also made of the effect of the chain length of the detergent on its interaction with thyroglobulin. Methods and Materials A Beckman Model GS pH meter, equipped with external electrodes, was used for all pH measurements. Water from a constant temperature bath flowed through a water jacket surrounding the titration cell. Xitrogen was blown over the top of the solution when measurements were made above pH 7. Titrants were added from a Rehberg microburet of 0.20-ml. capacity which was calibrated in 0.001 ml. units. All pH changes were complete in the time required to read the pH meter. The pH meter was calibrated with buffers recommended by Bates, el aZ.* The preparation and properties of thyroglobulin extracted from calf thyroid tissue have been reported in detail in
-
(1) J. Steinhardt, C. H. Fugitt and M.Harris, J . Research Natl. Bur. Standards, 24, 335 (1940); 25, 519 (1940); J. Steinhardt. Ann. N . Y . Acad. Sci., 41, 287 (1941). (2) F. W. Putnam and H. Neurath, J . Biol. Chem., lS9, 195 (1945). (3) G. Scatchard and E. S. Black, THISJOURNAL, 68, 88 (1949). (4) R. G. Bates, G. D. Pinching and E. R. Smith, J . Research Natl. BUT.Standards. 45, 418 (1950).
3
I 4
1
I
5
6
7 OH
8
9
IO
11
Fig. 2.-The effect of 0.015 M SDS on the titration curve of thyroglobulin in 0.01 M KN08. Neutral solutions of thyroglobulin were titrated with either 0.097 M HCl or KaOH. earlier papers.536 A phosphate-fractionated preparation ( l ) 5 was used in the observations currently reported. The sulfate series of alkyl detergents were purified preparations and were a gift from Dr. E. Barthel of E. I. du Pont de Nemours & Co. Other reagents were C.P. grade or equivalent. Fresh glass distilled water was used throughout.
Experimental Results When small amounts of SDS were added to a thyroglobulin solution below pH 7.5, the pH of the solution increased. If the solution then was returned to its original pH with HC1, a second aliquot of SDS may be added and another rise in pH noted. This procedure then may be continued until the increase in pH elicited by the detergent becomes negligible. I n this manner a curve can be constructed which shows the influence of detergent on t’he hydrogen ion binding properties of thyroglobulin. Illustrative H+ binding curves, obtained a t several pH values, appear in Fig. 1.’ Above the isoelectric point of thyroglobulin,8 ie., pH 4.5, all the curves are of the Langmuir type; below pH 4.5 the curves exhibit an inflection point which re-
-
( 5 ) H. Edelhoch, J . Biol. Chsm., 288, 1326 (1960).
(6) H.Edelhoch and R. E. Lippoldt, ibid., 236, 1335 (1960). (7) Since the magnitude of these curves is strongly dependent on the ionic strength of the solution the pH changes are attributed to electrostatic effects and not to rupture of hydrogen bonds. Moreover at pH 5.4 in 0.01 M KNOa the addition of urea up to a final concentration of 5 M produced only a small increase in pH; similarly, guanidinium chloride up to 2 M produced only a small decrease in pH. (8) bf. Heidelberger and K. 0. Pedersen, J . Gen. Physiol., 19, 95 (1935).
NOTES
1772
Vol. 64
0.015 M SDS (in 0.01 M KN03) thyroglobulin is almost completely dissociated and possesses an effective volume many times that of UC. Interpretation of the effects of SDS on the titration curve of thyroglobulin may be attempted in terms of the influence of the binding of anionic detergent on the charge and configuration of the protein molecule. For this purpose the model used by Linderstrdm-Lang'O to evaluate the electrostatic effects on the titration curves of proteins will be used; however, w and R are considered as empirical electrostatic and configurational parameters 0.2
0.4
0.6
ML.SDS.
pH
Ti - log ni- = pKi Ti
- 0.868~2
(1)
where
Fig. 3-The influence of detergent chain length on the bindmg of hydrogen ions to thyroglobulin at pH 5.20: KNOs = 0.01 M ; T = 28.1 . Detergents were the sodium salts of: octyl sulfate (C8); decyl sulfate (Clo); where ni is the total number of each type of group dodecyl d f a t e (CH) and tetradecyl sulfate (G); all were having an intrinsic dissociation constant pKi; Ti in 0.10 M solution.
sembles a cooperative interaction. This type of curve could result if an increase in the number of detergent binding sites occurred as a result of detergent binding. To facilitate evaluation of the curves described above potentiometric titrations of thyroglobulin, with and without SDS, were obtained. The solid lines in Fig. 2 represent the pH values observed after adding HC1 or NaOH to a solution of thyroThe open and filled circles globulin at pH -7. indicate the pH values obtained on back-titrating acidified solutions of thyroglobulin with and without SDS, respectively. It can be seen in Fig. 2 that the back-titration curve without detergent deviates almost immediately from the forward curve. In SDS, however, the reverse curve coincides with the forward curve from pH 2.7 to 5.2. Above pH -7.5 the alkaline segment of the titration curve was almost independent of the presence of detergent ions. A small displacement was evident above pH 10 (cf. Fig. 2). Karuah and Sonenbergghave shown that the free energies and enthalpies of binding of alkyl sulfates to serum albumin are decidedly dependent on chain length between 8 and 12 carbon chains. We have observed a similar influence of chain length as evidenced by the H + binding curves reproduced in Fig. 3. A marked effect of chain length is apparent between the 8 and 12 unit chains. Apparently the effectiveness of the long chain approaches a limit near the dodecyl size since tetradecyl sulfate interacts only slightly stronger. In fact as shown elsewhere6 the tetradecyl sulfate produces less dissociation and unfolding of thyroglobulin than does decyl sulfate. The dependence of binding on chain length suggests a van der Waals component to the binding energy. Discussion It has been reporteda that SDS a t concentrations below 0.001 M dissociates thyroglobulin into halves without much effect on their shape; however, a t higher detergent concentrations both thyroglobulin and its subunit swell or unfold appreciably. In (9) F. Karush and M. Sonenberg, J . Am. Cham. Soc., 71, 1369 (1949).
is the number of groups dissociated a t a given pH; w is the electrostatic work required to remove a proton from a protein molecule which bears the net charge 2 and has an effective radius of R; A is the radius of exclusion ( A = R 2.5)) K is the reciprocal Debye radius, D is the dielectric constant, h: is the Boltzmann constant and E is the electronic charge. Near the isoelectric point (pH 4.5) the binding of detergent anions results in the protein acquiring a net negative charge. In accordance with equation l an increase in the affinity of proton binding groups results and as a consequence an absorption of protons coupled with a rise in the pH of the solution. Acid to the isoelectric point, the positive charge enhances the protein's affinity for detergent anions. The large shift in the titration curve below pH 6 is illustrated in Fig. 2. However, in the acid region, the number of carboxyl groups which can accept protons is progressively reduced. Hence as the SDS titration is performed a t lower pH values fewer dissociated carboxyl groups remain thereby accounting for the peak in the H+ binding curves shown in the inset in Fig. 1. This curve may be obtained also from the vertical differences between the two titration curves shown in Fig. 2. In addition to the effect of SDS binding on the net charge (2) configurational changes may also be induced which will affect the value of R and therefore the electrostatic interaction term w. On the alkaline side of the isoelectric point the binding of detergent ions would be expected to decrease as the protein (negative) charge increased. The difference between the ttr-o thyroglobulin titration curves above pH -7 in Fig. 2 is quite small. Thyroglobulin in 0.01 M SDS is more extensively unfolded than the 5-15 and 5-12 components found below pH 11 in the absence of detergent.1l The greater degree of unfolding and consequently larger effective value of R would lead to a smaller value of w in equation 2; however the
+
-
(10) K. Linderatr6m-Lang, Compt. rend. h e . Lab. Cadsbeto, 16. No. 7 (1924). (11) 5-16 and 8-12 refer t o molecules with Svedberg sedimentation constants close to 15 and 12, respectively, which are formed from thyroglobulin (S-19) by raising the pH ofits solution.
Kov., 1960 larger value of Z which results from detergent binding appears to compensate so as to leave the net effect (wZ) approximately the same as observed without SDS. When thyroglobulin is titrated with HC1 in 0.015 M SDS, the titration of the carboxyl groups is more than 90% complete at pH 4.0. Moreover, thyroglobulin remains soluble at all pH values in 0.015 M SDS whereas it is insoluble between -4.6 and 3.6 without detergent. Without SDS less than 50% of the carboxyl groups are titrated at pH 4.0 (see Fig. 2). In the latter instance, the pH must be reduced to below 3 to reach comparable degrees of H + binding. It is evident therefore that potentiometric titration in SDS offers a convenient way of obtaining- the maximum number of acid bindinggroups.12 It is known that thvrodobulin is readilv denatured by reducing the bHXo the vicinity of its isoelectric point, ie., 4.5.8 As seen in Fig. 2 the difference between the forward and reverse titration curves increases gradually from pH 2.9 to 7.0. When SDS is present, the reverse curve coincides with the forward curve from pH 2.7 to 5.2 and then shifts toward the alkali region. The implication of these experiments is that whatever transformation is occurring in acid solution, it is still incomplete a t -pH 3.0 without SDS, while with detergent present the transformation is complete by pH 5.2. It is apparent, therefore, that SDS can greatly facilitate the disorganization of thyroglobulin to a more elementary form in its structure. (12) When bovine serum albumin and ovalbumin were titrated in SDS with HC1 from neutrality, the forward (acid) and reverse (base) curve were essentially identical and almost all the carboxyl groups were titrated by pH 4.0. I t is evident that SDS has the property of shifting the titration range of carboxyl groups in proteins to higher pH values. hforeover, with many insoluble proteins it is possible to keep the protein in solution with detergent. Finally, if acid binding groups are present in the protein which are "masked" or unavailable under nornial titration procedures, these groups then may become accessible l o titration.
NOTES
1773
15.0 1
F; 17.0
I 18.0-
19.0 --
t
20.0 i
0.0031 0.0032 0.0033 1/T, "K.-1. Fig. 1.-A plot of In (k'/T) us. 1/T for the racemization of GG-nitro-2,2'-carboxybiphenyl a t one atmosphere pressure. 0.0029
0.0030
ward interpretation. Since such data are apparently not available, a study of the racemization of the sodium salt of &6-nitro-2,2'-carboxybiphenyl at pressures from atmospheric to 12,000 atmospheres was made. Over the past 30 years, numerous diphenyl derivatives have been prepared and resolved4 and it is well known that the optical activity is the result of hindered rotation around the bridging bond. In this study 1-6-nitro-2,2'-carboxybiphenyl was chosen because of the relative ease of synthesis and the convenient temperature range in which it racemizes at a measurable rate.j
Experimental Procedure For the measurement of the optical rotation of the samples, a Landolt-Lippich Triple Shadow Polarimeter was employed. The polarimeter was cslihrated to f 0.01", .but because of the strong color (oamber) of the sample solution, readings were made to f0.1 BY C. C. MCCUNE,F. WM. CAGLE,JR., AND S. S. KISTLER The E-6-nitro-2,2'-carboxybiphenyl had a specific rotation of &,6, -280". Department8 of Ctiemzcal Engineerzng and Chemzstry. Unzversily o/ Utah, Salt Lake City 18. Utah Approximately40 mg. was used in each test, dissolved in 6 ml. of 0.10N NaOH and viewed in semi-micro polarimeter Received M a y b f , 1980 tubes The initial rotation was about 4' when the mercury It was desired to investigate the effect of hydro- green line, 5461 was used. For the transmission of this green line, Wratten filters static pressure on the velocity of a particularly simple chemical reaction. For such a study, one number 77a and number 58 were used. Filter number 77a transmits 68% of the green line with a negligible per should avoid especially reactions which might cent. of the yellow line. Filter number 58 eliminates the
THE EFFECT OF HYDROSTATIC PRESSURE ON THE RATE OF RACEMIZATION OF 1-GNITRO-8,2'-CARBOXYBIPHENY L
.
w.,
involve appreciable differences in ionic character in the initial and activated states, since interpretation of these is complicated by possible solvation The pressure effect on the rate of racemization of an optically active biphenyl should provide a particularly simple reaction, and the data obtained should be open to straightfor-
(1) J. Buchanan and S. D. Hamann. Trans. Faraday Soc., 49, 1425 (1953). (2) 9. D. Hamann, "Physico-Chemlcal Effects of Pressure," London, Butterworths Scientific Publications, 1957, Chap. 9. (3) C. T. Burris, Ph.D. Thesis, Catholic University, Washington, D. C., 1955.
k'
red line.
The rate constant for the reaction I
then could be obtained from 2k't = In
%-
d
(o(~/cY)
in which k' is the rate constant in sec.-1, t is the time in seconds, cro is the initial rotation, and a is the rotation at timet. A plot of ln(ki'/T)vs. 1/Tin which T i s the absolute temperature and k,' is the rate constant a t one atmosphere, allows one to determine the enthalpy and entropy of attiR. Adams and H. C . Yuan, Chem. Reus., 12, 261 (1933). ( 5 ) F. Bell and P. H. Robinson, J . Chein. Soc., 2234 (1927). (4)