Thermodynamics of Ionization of Amino Acids Part 6.t-The Second Ionization Constants of Some Glycine Peptides BY (THE LATE) EDWARD J. KING Department of Chemistry, Barnard College, Columbia University, New York, U.S.A. Received 30th May, 1974 Thermodynamic functions for the ionization of the ammonium groups of glycylglycine, glycyl-DL- alanine, glycyl-DL-a-amino-n-butyric acid, and glycyl-L-leucine were obtained from e.m.f. measure- ments at temperatures from 5 to 50°C on cells without liquid junction. The ionization functions of the carboxyl groups of six glycine peptides have been The ionization of the ammonium group reported in an earlier paper of this series.l +NHSCH2CONHCH(R)COO-+ HzO+NH2CH,CONHCH(R)COO-+ H30f has now been investigated for four peptides, namely those in which the substituent group R is hydrogen, methyl, ethyl or isobutyl.The ionization constants of the ammonium group have the form K2 = [H,O+][A-]/[HA*] where HA* represents the dipolar form of the peptide and A- the anionic form. The second ionization constants were obtained from measurements of the e.m.f. of cells of the type Pt, H,IHA* (mJ, NaA (m,), NaCl (m,)lAgCl, Ag (1) where m,, m2 and m3 are the stoichiometric molalities. Twenty-five buffer solutions were used for determination of the constants of glycylglycine. These were divided into three groups having stoichiometric buffer ratios, p = m,/m,, of approximately 2, 1, and 0.5. With each of the other peptides eight buffer solutions, all with p = 1, were used. Measurements were made at ten temperatures from 5 to 50°C in order to obtain accurate values of the enthalpies and entropies of ionization.EXPERIMENTAL Glycylglycine, glycylalanine, and glycyl-a-amino-n-butyric acid were the same prepara- tions as used in the earlier w0rk.l However, some of the earIier supply of glycyl-L-leucine [2-(N-glycyl)amino-4-methylpentanoic acid] was supplemented by a new sample of glycine- free material from the same source (H.M. Chemical Co., Santa Monica, California). The new material was found to be 99.94 % pure by form01 titration and contained less than 0.004 % each of ammonia, chloride, phosphate, iron and heavy metals. Sodium chloride containing less than 0.001 % bromide was prepared and fused according to the directions of Pinching and Bates.2 Carbonate-free sodium hydroxide solution was prepared and standardized against National Bureau of Standards potassium acid phthalate as in earlier Stock -f Part 5, E.J. King, J. Amer. Chem. SOC., 1960, 82, 3575. 88E. J. KING 89 buffer solutions were made from weighed amounts of peptide, sodium chloride, sodium hydroxide and water. These stock solutions were diluted further to give the cell solutions which were boiled under reduced pressure and swept with hydrogen. Buoyancy correction for air or hydrogen was made as required. During preparation of the solutions and filling of the cells, exposure of the solutions to air was kept to a minimum to avoid contamination but none was found. The test was carried out by mixing about 5 cm3 of buffer solution with 3 cm3 each of 1.0 mol dnr3 barium chloride solution and 0.2 mol dm-3 carbonate-free, sodium hydroxide solution.The appearance of the mixture after 5-10 min was compared with turbidities developed from known amounts of carbonate. Solutions were viewed in a blackened comparator block and as little as 3 x mol dm-3 could sometimes be detected though the limit of detection was usually about twice this. Measurements over the temperature range required about ten hours and were made in the following sequence : 25, 20, 15, 10, 5, 30, 35, 40,45,50 and 25°C. An indication of the performance of the cells is given by comparing the initial and h a 1 readings at 25°C. The standard deviations between initial and final values were 0.12 mV for glycylalanine, 0.063 mV for glycyl-a-amino-n-butyric acid, and 0.065 mV for glycylleucine.For glycylglycine buffers with ionic strengths above 0.02 mol kg-' the standard deviation was 0.12 mV. The performance of cells containing more dilute glycyl- glycine buffers was generally much poorer, the standard deviation in EZ5 for such cells being 0.21 mV. A few results obtained from these dilute buffers appeared to be low and were neglected, as discussed later, in calculating the ionization constant of glycylglycine. Smoothed values of the e.m.f. corrected to 1 atm hydrogen gas partial pressure are given as parameters of the equation, where t is in "C, The apparatus has been described Et = E25+ ~(t-25)+ b(t- 25)' in table 1. The standard deviations between observed and calculated values were & 0.085 mV for glycylglycine, f 0.035 mV for glycylalanine, & 0.056 mV for glycylaminobutyric acid, and k0.026 mV for glycylleucine.Actual experimental values were used in the calculations. TABLE PARAMETERS OF THE EQUATION E' = EZ5+ a(t- 25)+ b(t- 25)' 10 3ml/moi kg-1 2.572 4.967 6.185 7.226 10.186 15.107 19.34 30.84 5.270 8.243 9.638 12.131 15.819 18.560 23.60 32.58 40.51 53.56 103ma/mol kg-' lo3 m3/mol kg-' E25/V glycylglycine ( p = 4, m1-4m2-4m3) 5.028 9.710 12.240 14.126 20.16 29.90 3 8.27 61.03 5.067 9.787 12.283 14.237 20.23 30.00 38.41 61.25 0.863 98 0.847 61 0,842 71 0.838 37 0.830 29 0.820 69 0.814 83 0.803 97 glycylglycine ( p = 1, ml = m2 = m3) 5.288 8.273 10.005 11 -893 15.507 18.627 23.68 32.58 40.51 53.56 5.286 8.269 10.004 12.01 3 15.664 18.618 23.67 32.63 40.56 53.63 0.846 17 0.835 12 0.831 17 0.825 37 0.818 73 0.815 10 0.809 27 0.801 51 0.796 37 0.789 89 - 10k/V K-' - 10Eb/V K-' -5 + 52 73 85 116 148 172 21 3 59 96 110 133 154 166 188 216 235 258 245 235 245 240 245 250 245 250 245 240 235 235 240 240 240 240 240 24090 IONIZATION OF AMINO ACIDS 103m1/mol kg-1 10.566 18.99 40.29 51.32 68.28 89.29 109.00 6.880 10.462 15.322 20.21 24.25 29.44 34.55 39.60 10.272 11.041 16.256 20.05 23.76 29.53 35.46 40.54 7.986 10.191 15.004 17.98 24.48 29.88 34.83 40.19 TABLE 1-continued 103rn2/mol kg-1 lo3 m3/mOl kg-1 E2 5/v - 106a/V K-' - 106b/V K-* glycylglycine (p = 2, ml = 2m2 = 2m3) 5.176 5.212 0.828 24 9.301 9.708 0.812 91 19.74 20.60 0.794 54 25.14 26.24 0.788 77 33.63 43.09 0.776 86 43.98 56.34 0.770 72 53.69 68.78 0.766 26 glycyl-~~-alanine (p = 1) 6.938 10,519 15.453 20.32 24.38 29.50 34.63 39.69 6.909 10.490 15.387 20.26 24.32 29.47 34.59 39.65 0.843 82 0.833 33 0.824 07 0.817 15 0.812 74 0.808 09 0.814 25 0.801 02 glycyl-DL-a-amino-n-butyric acid (p = 1) 10.235 11.086 16.320 19.98 23.67 29.26 35.43 40.1 6 7.913 10.070 14.867 17.82 24.19 29.4 1 34.28 39.56 10.252 11.062 16.286 20.02 23.71 29.39 35.29 40.35 0.833 59 0.831 89 0.822 36 0.817 23 0,813 03 0.807 72 0.803 34 0.800 15 glycyl-L-leucine ( p = 1) 7.950 10.129 14.936 17.90 24.33 29.64 34.56 39.87 0.839 55 0.833 51 0.823 95 0.819 62 0.812 01 0.807 15 0.803 48 0.800 09 113 176 236 261 298 327 344 110 146 178 202 220 236 248 26 1 154 161 192 209 226 244 258 27 1 115 140 170 185 21 5 23 1 244 256 250 255 240 240 240 225 225 236 225 248 225 225 235 240 230 245 225 225 245 250 240 250 240 235 258 235 245 235 245 245 245 CALCULATIONS AND RESULTS From the experimental data the quantity -log YYIHYHYC~, can be calculated from -log mHyHyC1- = (E - E")/k +log m3 (1) where k = (RT/F)ln 10 and E" is the standard e.m.f.s* This function is related to PK2 byE.J. KING 91 In eqn (2) the hydroxyl ion concentration, m&, which is very small in comparison with m, and m2 is given with sufficient accuracy by (3) The activity coefficient term in eqn (2) should depend on both the ionic strength (Z = m2 +m3) and the dipolar ion concentration (m,). Alternatively, the inde- pendent variables can be chosen as Z and p = m /m2 since Z is proportional to m, for the buffer solutions used in this investigation. If the behaviour of glycine+ potassium chloride solutions is used as a guide, the activity coefficient term in eqn (2) should be expressed by -log mbH = pK, +log mHyH.and taurine +hydrochloric acid solutions (4) Y C l - log ---?HA * = (a + a’p)l+ (p f p’p)12 + (6p + 8’p2)I2 YA - where the Greek letters represent constants. In dilute solutions only the first term on the right-hand side of eqn (4) should be important, with the result that A plot of pKi against Z will thus give pK2 at Z = 0 and have a slope which varies with the buffer ratio. Such graphs for glycylglycine at 25°C are illustrated in fig. 1 and those for the other peptides in fig. 2. If the unreliable results below Z = 0.02 are discounted, the plots are straight lines up to Z of about 0.06. The curvature at higher pK$ pK2+(a+a’p)Z.(5) 8.34 I I I I I 0 0 O / 4 8.30 - -- 0 /e O 0 826 - I I I I I 20 40 60 80 I00 103ml /mol kg-’ FIG. 1 .-Extrapolation of glycylglycine results at 25°C against acid molality (ml) according to eqn(6): $ , p = & ; @ , p = l ; @ , p = l ; O , p = 2 .92 IONIZATION OF AMINO ACIDS concentrations is to be expected because of neglect of higher terms of eqn (4). Alter- natively, one could choose a in eqn ( 5 ) by trial so that points of a plot of (pK5 - aZ) against m, all fall on the same curve irrespective of the buffer ratio. At 25°C for glycylglycine, the expression fits all the data except the three points at high ionic strengths. pK; -0.3801 = pK, + f ( m l ) (6) I I I I I I I t 8.28 I I I I I 40 6 0 80 1 0 0 I 2 0 8.24 x ) 1 031/mol kg-l FIG.2.-Extrapolation of glycylglycine results at 25°C using experimental values for activity co- efficients according to eqn (9) : o, p = 3 ; a, p = 1 ; 0, p = 2. A third method of treatment of the data for glycylglycine is possible utilizing data and in aqueous sodium * (NaC1). Combining these data, which are presented in the for the activity coefficient of glycylglycine YHAi chloride solutions,' form of polynomials, for the activity coefficient, then in water log YHA* (NaCl) - - -0.23296~~,+0.17976~~: -0.5216~~3 -0.3357mIms +0.7723~$ (7) YHA* (0) where terms in higher powers than rn: have not been included because they are negli- gible for the present purposes. Assuming log yA-/yCl- in eqn (2) is proportional to I, and any salting effect of A- on HA' is proportional to m, (and hence to I ) , it is therefore reasonable to define a new extrapolation function pK; -= pKL +log = pK2 +BI (8) YHA * (NaCl) YHA'(0) where B is an adjustable parameter.A plot of eqn (8) against Zfor the results at 25°C is shown in fig. 3. It may be noted that results for I = 0.02 remain low but a straight line can be drawn through all the remaining points for the three buffer ratios to give a value of pK, in close agreement with the value obtained using eqn (5).E. J. KING 93 in determining the ionization constants of glycine and its Datta and Grzybowski N-methyl derivatives used the relation This is similar to that derivable from eqn (4) except that B and C were not expressed as functions of the buffer ratio since Datta and Grzybowski assumed the activity co- efficient of the dipolar ions was unity.In applying this relation to their data and to the data of King for glycine they gave equal weight to all points including those at PK; = pK2-BI-CI%. (9) 8.36 8.34 8.32 8.3, 8.3 2 0 40 60 80 1 031/mol kg- FIG. 3.-Extrapolation of results at 25°C for glycylalanine (0), glycyl-a-amino-n-butyric acid ( 0 ) and glycylleucine (0). the lowest ionic strengths. Since pK, values at ionic strengths below 0.02 tend to be low, to include these points makes the extrapolations show pronounced curvatures at low ionic strengths instead of the simple linear behaviour. If the lowest point in their extrapolations (in their fig. 1) is ignored, good straight lines could be drawn through the rest of the points up to I = 0.05.Above this ionic strength, the lines curve over as do those for glycylglycine (fig. 1). This is demonstrated using extra- polation based on eqn (5) in fig. 4. The failure to obtain useful results from cells containing very dilute buffers may be94 IONIZATION OF AMINO ACIDS caused by a number of sources of error which are trivial for more concentrated solu- tions. The silver-silver chloride electrodes are porous and may trap water or oxygen which are not removed by mixing when the cells are filled. Traces of carbonate could alter the buffer ratio. At the pH values of these buffers about 99 % of the carbonate would be converted to hydrogen carbonate ion. Then mHA* = m1-k mbH + mHc0; m A - = nz2 - mbH - mHCOj and the term log[(m, +mbH)/(m2 - m 3 ] in eqn (2) would be too small, making pK4 too small.Yet, to account for the errors in pK4 in very dilute solutions would require about 5 x mol dm-3 HCO:, and even one-tenth of this could not be detected in the cell solutions. It is readily shown also that corrections for the solu- bility of silver chloride in the buffer solutions l * l 2 are entirely negligible. Another possibility would be the malfunctioning of the hydrogen gas electrode in these very dilute buffer solutions of poor buffer capacity. Further investigations would be necessary to decide whether the effect is a real one, or an artefact of the electrodes in these solutions. 1 I I I 20 40 6 0 80 1031/mol kg-' x , ref. (11). FIG. 4.-Comparison of extrapolation of results for glycine : 0, ref. (4) ( p > 1) ; @, ref.(4) ( p < 1) ; The values of pK2 obtained by graphical extrapolation are summarized in table 2. Those for glycylglycine are the averages of the three values obtained by independent extrapolations of results for the three buffer ratios (fig. 1). The deviations in pKz listed in table 2 give an indication of the average precision of the extrapolations. Table 2 also includes the parameters A, B, and G of the Harned-Robinson equation l3 pK, = (A/T)-B+CT. (10)E. J. KING 95 From these parameters the changes in various thermodynamic properties for the ionization reaction in the standard state, can be calculated. Values at 25°C are given in table 3 for AGq, the change in free energy; AH;, the change in enthalpy or the heat of ionization; AS;, the change in entropy.TABLE 2.-vALUES OF pK2 FOR GLYCINE PEPTIDES AND THE PARAMETERS OF EQN (8) glycyl-a-amino-n- temp./K glycylglycine glycylalanine butyric acid glycy'leucine 278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 dev. AIK B C1K-l 8.813 8.668 8.529 8.394 8.265 8.140 8.018 7.901 7.788 7.680 k0.002 2 412.46 0.316 3 0.001 6420 8.896 8.746 8.602 8.462 8.331 8.202 8.077 7.958 7.843 7.732 f 0.001 2 580.60 1.168 2 0.002 828 4 8.899 8.748 8.604 8.464 8.331 8.200 8.073 7.952 7.837 7.725 + 0.002 2 563.10 0.985 6 0.002 410 1 8.887 8.740 8.597 8.459 8.327 8.199 8.074 7.955 7.840 7.729 *O.Ool 2 452.31 0.354 3 0.001 5298 The values of the ionization constants can be compared with the few values available in the literature. Neuberger l4 using cells with liquid junction and a double extrapolation, obtained for glycylglycine pK2 = 8.265 at 25"C, in precise agreement with that recorded in table 2.Smith and Smith l 5 obtained a lower value, pK2 = 8.255. They used the same method as the present investigation, but they measured the electromotive forces of only five cells to the nearest 0.1 mV and did not exclude oxygen from the solutions. Enthalpies and entropies of ionization derived from their data are also not in agreement with those reported here. Of numerous less precise determinations of ionization constants by potentiometric titration that of Monk is typical. The only values His value of pK2 for glycylglycine at 25°C is 8.23 4 0.01. TABLE 3.-THERMODYNAMIC PROPERTIES FOR THE IONIZATION OF NH3CH2CONHCH(R)COO- AT 298.15 K AG"/kcal mol-1 AH"/kcal mol-1 AS"/cal mol-1 K-1 glycine N-gl ycylgl ycine alanine 21 N-gl ycyl alan r ne a-amino-n-butyric acid N-gl ycyl-a-amino-n- bu t yric acid leucine 21 N-gl yc ylleucine N-methylglycine NN-dimethylglycine bicine a* 22 tricine b* 22 13 338 11 275 13 460 11 365 13 410 11 364 13 295 11 360 13 916 13 561 11 369 11 099 10 550 10 370 10 980 10 660 10 696 10 750 10 815 10 600 9 681 7 654 6 279 7 520 - 9.36 - 3.03 - 8.32 -2.37 -9.10 - 2.07 - 8.32 -2.55 - 14.20 - 19.81 - 17.07 - 12.00 u bicine is NN-di-(2-hydroxyethy1)glycine ; b tricine is N-tris(hydr oxymethy1)methylglycine.96 IONIZATION OF AMINO ACIDS available for glycylalanine and glycylleucine pK, were also determined by potentio- metric titration.Perkins l7 reports 8.21 and 8.17 respectively at 25°C.Ellenbogen l8 found 8.23 for glycylalanine. Salakhutdinov et aL1’ give 8.28k0.01 at 25°C and Dobbie and Kermack 2o report 8.41 at 20°C for glycylleucine. All of these values are lower than those reported in table 2. Table 3 gives a comparison of AG;, AH;, and AS; at 25°C for each of the four N-glycyl substituted amino acids and corresponding data for the parent unsubstituted amino acids.4* 21 The effect of N-glycyl substitution on AGZ is to lower it by between 1935 and 2095 cal mol-l, 2035 cal mol-l on average. The effect on AH; is much smaller with an average diminution of 165 cal mo1-l with a small increase of 54 cal mo1-1 for N-glycyl-a-amino-n-butyric acid. Consequently, changes in AS; on N-glycyl substitution depend mainly on changes in AC;.The average value of AS; is - 8.78 cal mol-1 K-I for the substituted acids representing a significant difference. The effect of N-glycyl substitution is almost independent of the nature of the R substituent for the four amino acids now investigated. However, with R = H, much more variable effects are found for substituents other than N-glycyl. Table 3 illustrates this for four different substituents.ll9 22* 23 There is some variation between the AG; values but very marked variations in the AH; values and consequently the AS; values. The experimental work was initiated at Barnard College in 1959 supported by a research grant from the National Heart Institute of the National Institute of Health, U.S. Public Health Service. This paper was prepared from Prof.King’s incomplete manuscript and notes by Dr. A. K. Covington in collaboration with Dr. Grace W. King and Prof. R. A. Robinson of the Department of Physical Chemistry, University of Newcastle upon Tyne, United Kingdom, where Prof. King was spending study leave at the time of his death. E. J. King, J. Amer. Chem. SOC., 1957, 79, 6151. G. Pinching and R. G. Bates, J. Res. Nat. Bur. Stand., 1946, 37, 311. E. J. King, J. Amer. Chem. SOC., 1954, 76, 1006. E. J. King and G. W. King, J. Amer. Chem. SOC., 1956,78,1089. E. J. King, J. Amer. Chem. SOC., 1956, 78, 6020. E. J. King, J. Amer. Chem. SOC., 1953, 75, 2204. H. D. Ellerton, G. Reinfelds, D. E. Mulcahy and P. J. Dunlop, J. Phys. Chem., 1964, 68,398. 4E. J. King, J. Amer. Chem. SOC., 1951, 73, 155. ’ R. M. Roberts and J. G. Kirkwood, J. Amer. Chem. SOC., 1941, 63, 1373. lo E. E. Schrier and R. A. Robinson, J. Biol. Chem., 1971, 246, 1179. l 1 S. P. Datta and A. K. Grzybowski, Trans. Faraduy SOC., 1958, 54, 1179. l2 C. B. Monk, Trans. Faraday SOC., 1951, 47,292. l3 H. S. Harned and R. A. Robinson, Trans. Faraday SOC., 1940, 36, 973. l4 A. Neuberger, Proc. Roy. SOC. A, 1937, 158, 68. l5 E. R. B. Smith and P. K. Smith, J. Biol. Chem., 1942, 146, 157. l 6 C. B. Monk, Trans. Faraday SOC., 1951, 47, 285. l7 D. J. Perkins, Biochem. J., 1954, 57, 702. l8 E. Ellenbogen, J. Amer. Chem. SOC., 1952, 74, 5198. l9 U. I. Salakhutdinov, A. P. Borisova, Y. V. Granovskii, I. A. Savich and V. I. Spitsyn, Dokludy 2o H. Dobbie and W. 0. Kermack, Biochem. J., 1955,59,246. 21 P. K. Smith, A. C. Taylor and E. R. B. Smith, J. Biol. Chem., 1937, 122, 109. 22 S. P. Datta, A. K. Grzybowski and R. G . Bates, J. Phys. Chem., 1962, 68, 275. 23 R. N. Roy, R. A. Robinson and R. G. Bates, J. Amer. Chem. SOC., 1973, 95,2831. Akad. Nauk S.S.S.R., 1967, 177, 365.