Polymer Journal, Vol. 17, No. 4, pp 621-631 (1985)
Random-Coiled Conformation of Polypeptide Chains II. Experimentally Evaluated Characteristic Ratio of Poly(N5-2-hydroxyethyl• L-glutamine) and Theoretical Conformational Analysis of Poly(L-glutamine) and Poly(L-glutamic acid)
Masahito 0KA, Toshio HAYASHI*, and Akio NAKAJIMA
Department of Polymer Chemistry and *Research Center for Medical Polymers and Biomaterials, Kyoto University, Kyoto 606, Japan (Received September 12, 1984)
ABSTRACT: Random-coiled conformations of fractionated po1y(N 5 -2-hydroxyethyl-L-glu• tamine) were experimentally investigated and the characteristic ratio was obtained by the Stockmayer-Fixman equation with the experimental results for intrinsic viscosity and weight• averaged molecular weight. Moreover, random-coiled conformations of poly(L-glutamine) and poly(L-glutamic acid) were theoretically analyzed by conformational energy calculations based on intra-residue interactions, and the calculated characteristic ratio was in good agreement with the experimental results. KEY WORDS Random-Coiled Conformation I Characteristic Ratio I Intrinsic Viscosity I Ultracentrifugation I ECEPP 1 Conformational Energy Calculation I Poly(N 5-2-hydroxyethyl-L-glutamine) I Poly(L-glutamine) I Poly(L-glutamic acid) I
The characteristic ratio, a basic property of for normal synthetic polymers. 2) A random• random-coiled polymer chains in dilute solu• coiled conformation can exist only in a tions, provides information on local confor• limited solvent system16·17 in which intra- and mations of polymer chains. The values of side• inter-molecular hydrogen bonds of polypep• chain derivatives of poly(L-glutamic acid) and tides are broken. 3) The EJ-state14 of synthetic poly(L-lysine), treated as alanine-type poly• polypeptides has not been found yet because peptide chains, 1 •2 were experimentally investi• of helical conformations due to favorable gated. 3 - 13 The experimental values of 6 to I 0 short-range interactions such as backbone/ distribute over a somewhat wide range, due to backbone hydrogen bond; i.e., short-range experimental difficulties, polydispersity of the interactions are favorable compared to long• samples and uncertainty of the universal con• range interactions and polypeptidejsolvent in• stant 4J0 14·15 to calculate the characteristic teractions in the e-state. ratio from viscosity data. Experimental diffi• Such relations as the Stockmayer-Fixman, 18 culties are concerned with the following Orofino-Flory, 19·20 and Flory-Fox,21 used for points. 1) Synthetic polypeptides have the evaluation of the characteristic ratio from tendency to take helical conformations sta• viscosity data are derived for the mono• bilized by favorable intra-molecular hydrogen disperse polymer systems. However, many bonds and to cause molecular aggregations characteristics ratio3 ·7 ·9 · 10· 13 have been with intramolecular hydrogen bonds in evaluated from data on unfractionated sam• common solvents which are good solvents ples. Moreover, several values [(2.1-2.6) x
621 M. OKA, T. HAYASHI, and A. NAKAJIMA
1021 ] have been used as the universal constant introducing phosgene gas into suspensions of cf>0 14 in the Stockmayer-Fixman, Orofino• BLG in tetrahydrofuran (THF). BLG-NCA Flory, and Flory-Fox equations. This ambigui• was purified by repeated recrystallization from ty causes uncertainty in the characteristic ratio ethyl acetate solution with the addition of by as much as 5% to 20%. petroleum benzene. BLG-NCA was dissolved Using the fj-methylene group approxima• in the mixture of dry methylene dichloride and tion, the characteristic ratio of polypeptide dioxane (volume ratio I: 2), and polymerized chains composed of alanine-type residues were at room temperature in the presence of tri• theoretically investigated, 1.2.22 and 9.272 and ethylamine as the initiator ([M]/[I] =50). After 8.3822 were obtained. These values are polymerization for 3 days at room tempera• slightly higher than those experimentally ob• ture, poly(y-benzyl-L-glutamate) (PBLG) was tained.4-11 As shown in the previous paper,23 precipitated in excess cold methyl alcohol, a more precise analysis in which all intra• filtered and dried under reduced pressure. residue interactions including side-chain/ PBLG was aminolyzed through a reaction backbone interactions and freedom of internal with 2-amino-1-ethanol for 24 h at 55°C, 26 and rotations of the side chain are considered, dialyzed against five times the volume water of gives 11.24 and 12.33 for poly(L-phenyl• the reactant solution using a CU-8000 tube alanine) and poly(L-tyrosine), respectively, and with a Imp pore. After dialysis, poly(N5-2- the latter value was very close to that 12.324 of hydroxyethyl-L-glutamine) (PHEG) was pre• poly( 0-benzyloxycarbonyl-L-tyrosine ). These cipitated from an aqueous solution into ace• results indicate that side-chain/backbone inter• tone. The precipitated PHEG sample was actions are very important to stabilize local separated into 6 fractions with the water• informations of polypeptide chains even dioxane system, and each was freeze-dried. though they are not composed of fj-branched Completion of aminolysis was confirmed by residues. It should be of interest to make a the disappearence of absorption spectra of precise analysis of alanine-type polypeptide the benzyl-group at 257 nm, using a Hitachi chains such as poly(L-glutamine) and poly(L• Model ESP-3T Spectrometer. glutamic acid), considering the dependence of side-chain/backbone interactions on backbone Measurements conformations. Viscosity measurements were performed in In this paper, poly(N5-2-hydroxyethyl-L• 0.05 M sodium phosphate buffer at pH 7.4 and glutamine), which takes on a random-coil con• with an imposed Ubbelohde-type vis• formation12 in water at room temperature, was cometer having a flow time for the solvent synthesized as a model polypeptide of poly(L• longer than I 00 s. glutamine), and its characteristic ratio was The molecular weight of the samples was evaluated. Moreover, theoretical analyses of determined from equilibrium runs on polymer poly(L-glutamic acid) and poly(L-glutamine) solutions by a MOM Type-3170 B ultracen• were made, considering the side-chain/back• trifuge. As the solvent, 0.05 M sodium phos• bone interactions and freedom of i, i, and x3 phate buffer was used. Solution concentra• in the side-chain conformations. tions from 0.05 to 0.20 (g dl-1) were used in the sedimentation equilibrium runs. EXPERIMENTAL Optical rotatory dispersion (ORD) mea• surements were carried out with a JASCO J-20 Materials Recording Spectropolarimeter at 25oC. The The N-carboxy anhydrides (NCA) of y• concentration of the polymer solution was benzyl-L-glutamate (BLG) were prepared25 by 1.0 g dl-1 throughout the measurements to
622 Polymer J., Vol. 17, No.4, 1985 Random-Coiled Conformation of Polypeptides II. cancel out the effects of polymer concentration Glu-NHMe has 60 energy minima with AE < 3 on optical properties. The Moffit-Yang pa• kcal mol- 1 and all their x4 ·2 are almost 180° ± rameter -b0 was determined by the Moffit• 2o except that only two energy minima have Yang procedure.27 x4 ·2 = 24° (AE = 1.82 kcal mol- 1) and -26° (AE=2.14kcal mol- 1). Their results suggest THEORETICAL that fixing x4 of Gln and x4 ·2 of Glu at 180° has minor effects on the stability of backbone The nomenclature and conventions adopted conformations with intra-residue interactions, were those recommended by an IUPAC-IUB since the contributions of these conformations nomenclature commission.28 Assumptions and to the partition function (eq 1-1 33) are negli• definitions used in this work corresponded to gible. All other dihedral angles of backbone those in the previous work. 23 Conformational were fixed at 180°. energy E;(c/J;, t/1;, x;) of residue i was calculated for a model single residue peptide with two RESULTS blocking end groups, acetyl- and N-methyl• amide- (i.e., Ac-X-NHMe). All interactions in Experimental Evaluation of the Characteristic this model peptide are referred to as intra• Ratio residue interactions. The validity of this The weight-average molecular weight Mw, treatment is discussed in the paper 1. 23 degree of polydispersity Mz/ M w' intrinsic Conformational energy calculations were viscosity [ry], and Moffit-Yang parameter - b0 carried out for Ac-L-Gln-NHMe and Ac-L• are summarized in Table I. The degree of Glu-NHMe with ECEPP29 using the standard polydispersity Mz/ M w is relatively small, ex• geometry for bond lengths and bond angles, cept for the last two fractions, PHEG5 and except that the ca-H bond length was in• PHEG6. The value -b0 =0 indicates that PHEG takes on a random-coil conformation creased from 1.00 to 1.09 A. 30·31 The backbone conformations ¢ and t/J, and two side-chain in 0.05 M sodium-phosphate buffer at pH 7.4 and 25°C. Figure 1 shows a double logarithmic dihedral angles x1 and x2 were changed at !5° intervals. l was changed at 30° intervals, and plot of [ry] versus Mw for PHEG in 0.05 M sodium-phosphate buffer at pH 7.4 and 25°C. x4 of Gin and x4 ·2 of Glu were fixed at 180°. As shown by Zimmerman et al., 32 Ac-Gln-NHMe has 69 energy minima with AE < 3 kcal mol- 1 1-0 and all their l are nearly 180° ± 4°; also, Ac-L-
Table I. Molecular characterization of fractionated PHEG samples
[I]] Sample Mwx J0-4a M.fMw' code dl g-1 b
PHEGl 0.86 1.24 0.093 PHEG2 1.36 1.16 0.125 0-1 PHEG3 2.53 1.15 0.184 0 PHEG4 4.47 1.18 0.264 PHEG5 6.69 1.45 0.355 PHEG6 14.3 1.80 0.572 0 Figure I. Double-logarithmic plot of [17] vs. M w for ' Determined by a MOM ultracentrifuge 3170-b. PHEG in 0.05 M sodium phosphate buffer at pH 7.4 and b 0.05 M sodium phosphate buffer, pH 7.4, 25"C. 25'C.
Polymer J., Vol. 17, N0. 4, 1985 623 M. 0KA, T. HAYASHI, and A. NAKAJIMA
distance between successive a-carbon atom, n,
the degree of polymerization, and M 0 ( = 192), the molecular weight of monomer residue. In Figure 2, is plotted against Thus, g 1·5 0 'X K=0.865 x w- 3 0 2 2 0 2 3 4 5 As shown in paper I, the effects of side-chain - 112 conformations (x1 and on the stability of Mwx100 l) backbone conformations are very significant Figure 2. [1J]/Mw 112 plotted against Mw 112 for PHEG. (see Figure 4 in paper I). The side chains of L• Gln and L-Glu residues have two more free• The slope of the plot 0.65 also indicates that dom of internal rotation than those of the L• PHEG takes on a random-coil conformation Phe residue treated in paper I, i.e., X3 and x4 in this medium. The straight line in Figure 1 for L-Gln, and x3 and X4 '2 for L-Glu. As leads to the following Mark-Houwink• mentioned in Theoretical section, it is rea• Sakurada equation. sonable to fix x4 and l·2 at 180°. The 32 ['7]=2.62x theoretical results for single residue minima of L-Gln and L-Glu suggest that to the broad 0.86xl04 (4) 624 Polymer J., Vol. 17, No. 4, 1985 Random-Coiled Conformation of Polypeptides II. 0 60 120 180 cp ,... -... , I I 120 I I Iv I -60 f) () -120 0 60 120 180 cp Polymer J., Vol. 17, No.4, 1985 625 M. 0KA, T. HAYASHI, and A. NAKAJIMA 0 60 120 180 (c) 120 -60 -120 120 60 0 60 120 180 ¢ Figure 3. Energy contour ( ¢, t/J) maps of the L-Gln residue averaged over the (i, /) conformational sapce at 15 intervals. The contour lines are labeled as energy in kcal mol- 1 above the minimum energy point. The dashed lines indicate the 0. 5 kcal mol- 1 energy contour lines. 626 Polymer J., Vol. 17, No. 4,.1985 Random-Coiled Conformation of Polypeptides II. , ... -, I \ 1 I 120 I' '0 ' I \ ' ' 'J' 60 :=r-o J/ () -120 -120 -60 0 60 120 180 cp (a) l= -90' and xu= 180 60 J/ () Figure 4. Energy contour (¢, 1/J) maps of the L-Glu residue averaged over the (/, /) conformational sapce at 15° intervals. Polymer J., Vol. 17, No. 4, 1985 627 M. 0KA, T. HAYASHI, and A. NAKAJIMA Table II. Characteristic ratios of poly(L-gluta- 0.329 -0.112 0.654J mine) and poly(L-glutamic acid) for a specified [ side-chain dihedral angle l Poly(L-glutamine)' Poly(L-g1utamic acid)b 0. 764 -0.234 -0.251 l 2 ui 628 Polymer J., Vol. 17, No.4, 1985 Random-Coiled Conformation of Polypeptides II. 60 -60 [/ () -120 (a) f) () 0 60 120 180 (b) Figure 5. Energy contour(¢, 1/1) maps of L-Gin (a) and L-Giu (b) residues averaged over all side-chain conformations at !5° intervals for x' and/, 30° intervals for/, with l= 180° and x4 · 2 = 180°. Polymer J., Vol. 17, No.4, 1985 629 M. OKA, T. HAYASHI, and A. NAKAJIMA Table III. Experimentally evaluated cf>0 , and experimental difficulties. 8 of the characteristic ratios values were obtained for unfractionated sam• ples and show broad distribution in the Sample (R2 )o.x/nf2 Reference range 5.9 to 8.8. But three values obtained for PHEG 6.6 SP f This work the well fractionated samples were appro• 6.2 SF 12, 35 ximately the same as 8.28·37 and 7.7 11 for 10 OFFb ud 13 11 PBLG' 8.2 SF f 8 PBLG, and 7.7 for PMLG. The average 7.7 SF f II value was 7.9, and that of all 11 values from 7.5 SF u 6 Table III, 7.7. These values are close to the 5.9 SF u 3 theoretical results, 7.51, of poly(L-glutamic PMLG' 7.7 SF f II 7.7 SF u 10 acid) within experimental error. The theoreti• PELG' 6.9 SF u 7 cal results for poly(L-glutamine) and poly(L• PPLG" 6.2 SF u 9 glutamic acid) were obtained by exact calcu• PLG-Na; 8.8 OFF u 4 lations, considering all atom-atom pair inter• 6.5 OFF u 5 7.6 OFF u 5 actions in the backbone and side chain. How• ever, the atomic groups of the side-chain de• " Evaluated by the Stockmayer-Fixman equation. rivative were not considered in our theoreti• b Evaluated by the Orofino-Flory and Flory-Fox cal calculations. It is thus considered that equations. ' Fractionated samples. the experimental results should deviate from d U nfractionated samples. those theoretically obtained if additional in• ' Poly(y-benzyl-L-glutamate). teractions between substituent groups and ' Poly(y-methyl-L-glutamate). other residue parts are important. Fairly ' Poly( y-ethyl-L-glutamate ). " Poly(y-phenacyl-L-glutamate). good agreement was obtained between ex• ; Na-Poly(L-glutamic acid). perimental and theoretical results. More• over, there was no appreciable difference Table IV. Theoretically evaluated among the five substituents of poly(L-glutamic characteristic ratios acid). Thus group substitution in the side chain Polypeptide Oka eta/. Tanaka et a/. Flory eta/. may not have any significant effect on the characteristic ratios of poly(L-glutamine) and Gin 6.62 poly(L-glutamic acid). Glu 7.51 Ala 8.15" 8.38b 9.27' In our calculations, only the intra-residue Gly 2.15" 2.J7b 2.16' interactions were considered and not inter• Phe 11.24" residue interactions. Both of inter-residue Tyr 12.33" interactions, particularly side-chain/side-chain • From ref 23. interactions, and intra-residue interactions b From ref 22. were important for the conformational stabili• ' From ref 2. ty of oligopeptides.38 ·39 The good agreement between experimental and theoretical results As shown in Table III, the experimentally indicates that inter-residue interactions have evaluated characteristic ratio of various side• only minor effects on the characteristic ratios chain substituents of poly(L-glutamic acid) are of poly(L-glutamine) and poly(L-glutamic distributed in the range of 5.9 to 8.8. As acid). This is consistent with the calculation of mentioned in the Introduction section, this the characteristic ratio of poly(L-phenyl• wide distribution of characteristic ratios was alanine?3 and theoretical conformational due to the polydispersity of the polypeptide analysis of di- and tripeptides of L-phenyl• samples, ambiguity of the universal constant alanine.40 630 Polymer J., Vol. 17, No.4, 1985 Random-Coiled Conformation of Polypeptides II. Theoretical results summarized in Table IV (1975). indicate that the intra-residue side-chain/ 17. A. Teramoto and H. Fujita, J. Macromol. Sci., C, 15, backbone interactions of poly(L-glutamine) 165 (1976). 18. W. H. Stockmayer and M. Fixman, J. Polym. Sci., C, and poly(L-glutamic acid) lessen the charac• I, 137 (1963). teristic ratio but those of poly(L-phenyl• 19. T. A. Orofino and P. 1. Flory, J. Chern. Phys., 26, alanine) and poly(L-tyrosine) enhance it; that 1067 (1957). is, the overall stability of the backbone confor• 20. T. A. Orofino and P. 1. Flory, J. Phys. Chern., 63, 283 (1959). mation depends on side-chain groups beyond 21. P. 1. Flory and T. G. Fox, J. Am. Chern. Soc., 73, the /3-carbon atom. 1904 (1951 ). 22. S. Tanaka and A. Nakajima, Polym. J., 2, 717 (1971). Acknowledgements. We should like to 23. M. Oka and A. Nakajima, Polym. J., 16, 693 (1984). This paper is referenced as paper I. thank Professor A. Teramoto, Department of 24. 1. P. Vollmer and G. Spach, Biopolymers, 5, 337 Polymer Science, Osaka University, for pre• (1967). senting the experimental data of PHEG. 25. E. R. Blout and R. H. Karlson, J. Am. Chern. Soc., Computations were carried out by the 78, 941 (1956). 26. N. Lupu-Lotan, A. Yaron, A. Berger, and M. Sela, FACOM M-380 Computer at the Data Proc• Biopolymers, 3, 625 (1965). essing Center, Kyoto University. 27. W. Moffitt and 1. T. Yang, Proc. Nat/. Acad. Sci. U.S.A., 42, 596 (1956). 28. IUPAC-IUB Commission on Biological REFERENCES Nomenclature, Biochemistry, 9, 3471 (1970). 29. F. A. Momany, R. F. McGuire, A. W. Burgess, and I. D. A. Brant and P. 1. Flory, J. Am. Chern. Soc., 87, H. A. Scheraga, J. Phys. Chern., 79, 2361 (1975). 2791 (1965). 30. E. Benedetti, C. Pedone, C. Toniolo, G. Nemethy, 2. D. A. Brant, W. G. Miller, and P. 1. Flory, J. Mol. M. S. Pottle, and H. A. Scheraga, Int. J. Peptide Bioi., 23, 47 (1967). Protein Res., 16, !56 (1980). 3. P. Doty, 1. H. Bradbury, and A. M. Holtzer, J. Am. 31. G. Nemethy, M. S. Pottle, and H. A. Scheraga, J. Chern. Soc., 78, 947 (1956). Phys. Chern., 87, 1883 (1983). 4. D. A. Brant and P. 1. Flory, J. Am. Chern. Soc., 87, 32. S. S. Zimmerman, M.S. Pottle, G. Nemethy, and H. 2788 (1965). A. Scheraga, Macromolecules, 10, 1 (1977). 5. W. G. Miller, D. A. Brant, and P. 1. Flory, J. Mol. 33. Equation I in the previous paper (paper I) is abbre• Bioi., 23, 67 (1967). viated as eq I -1, and so on. 6. H. Fujita, A. Teramoto, T. Yamashita, K. Okita, 34. The value I= 3. 78 A of ECEPP is slightly smaller and S. Ikeda, Biopolymers, 4, 781 (1966). than the conventional value 3.8 A used in the expe• 7. M. Terbojevich, E. Peggion, A. Cosani, G. D'Este, rimental analysis. The effect of this difference on the and E. Scoffone, Eur. Polym. J., 3, 681 (1967). characteristic ratio being negligibly small (less than 8. T. Norisuye, thesis, Osaka University (1973). I%), the conventional value 3.8 A were used in 9. F. Heitz, E. Marchal, and G. Spach, Macro• experimental evaluation of (R2 ) 0 . x.lnfl by eq 3. molecules, 8, 145 (1975). 35. A. Teramoto, private communication. 10. S. Tanaka and A. Nakajima, Bull. Inst. Chern. Res., 36. a=0.35 is more compatible to Figure 2 of ref 13 than Kyoto Univ., 54, 229 (1976). a= 0.55 which mentioned in ref 13. II. G.-W. Chen, thesis, Kyoto University (1981). 37. The value 8.2 in Table III is recalculated from the 12. M. Miyake, S. Akita, A. Teramoto, T. Norisuye, and original data of ref 8 using <1>0 =2.5 x 1021 . H. Fujita, Biopolymers, 13, 1173 ( 1974). 38. I. D. Rae, S. 1. Leach, E. M. Minasian, 1. A. Smith, 13. W. L. Mattice and 1.-T. Lb, Macromolecules, 5, 734 S. S. Zimmerman, 1. A. Weigold, Z. I. Hodes, G. (1972). Nemethy, R. W. Woody, and H. A. Scheraga, Int. J. 14. P. 1. Flory, ""Principle of Polymer Chemistry," Peptide Protein Res., 17, 575 (1981). Cornell University Press, Ithaca, N.Y., 1953. 39. M. Oka, G. T. Montelione, and H. A. Scheraga, J. 15. H. Yamakawa, ""Modern Theory of Polymer Am. Chern. Soc., 106, 7959 (1984). Solutions," Harper & Row, New York, 1972. 40. M. Oka and A. Nakajima, to be submitted to 16. A. Teramoto and H. Fujita, Adv. Po/ym. Sci., 18, 65 Polym. J. Polymer 1., Vol. 17, No. 4, 1985 631