OPTICAL ROTATION AND INFRA-RED SPECTRA OF SOME POLYPEPTIDE AND PROTEIN FILMS BY A. ELLIOTT, W. E. HANBY AND B. R. MALCOLM Courtaulds Limited, Research Laboratory, Lower Cookham Road, Maidenhead, Berks Received 23rd January, 1958 Refinements of technique have enabled the dispersion of the optical rotation of solid films to be measured in the visible and near ultra-violet regions. Such measurements have been made for a series of polyalanines containing different proportions of D- and L- residues, for the sodium and potassium salts of poly-L-glutamic acid and also for some protein films. The infra-red spectra of these films have also been observed. The poly- alanine films show the characteristic dispersion of the a-helix, but the other materials do not. Since all these films have carbonyl absorption bands at ca.1660cm-1, it is evident that this frequency is associated with two or more configurations of the poly- peptide chain. With Bornbyx silk films cast from aqueous solution, a random coil appears possible. Infra-red spectra and X-ray diffraction patterns of polypeptides and fibrous proteins have chiefly been made on solid films, because the technique and inter- pretation is often more difficult if the material is in solution. On the other hand, measurements of optical rotation, which have recently yielded important informa- tion in connection with the a-helical configuration of polypeptides (Doty and Yang,l Moffitt and Yang,2 Elliott, Hanby and Malcolm,3 Yang and Doty,4 Moffitt, Fitts and Kirkwood,s Downie, Elliott, Hanby and Malcolm6) are much more easily made on liquids and recent published work has been restricted to solutions.Since the origins of the frequencies which characterize different polypeptide con- figurations have never received a satisfactory explanation and are empirical observations, it has seemed desirable to fill certain gaps in our knowledge by measuring the optical rotations of solid films on which observations of infra-red spectra and of X-ray diffraction patterns could be made. This has clarified a somewhat anomalous and unsatisfactory situation concerning the configuration of Bornbyx silk cast from an aqueous dispersion. EXPERIMENTAL METHODS FOR MEASURING OPTICAL ROTATION OF FILMS The peculiar difficulties which arise in measuring the optical rotation of films (as distinct from liquids) are caused by birefringence from strain, from local orientation of the polymer molecules or from birefringent foreign bodies in the film.Birefringence in the specimen produces an error in the se:ting of the analyzing polarizer and causes the field to be more or less brightly illuminated. This diminishes the sensitivity of a visual or a photo-electric polarimeter, since source fluctuations become troublesome when the field is bright. To prepare specimens which are sufficiently homogeneous, it is usually necessary to make a number of thin films rather than a few thick ones. In practice, reasonably accurate measurements can only be made with a suitable photo-electric polarimeter, and we have used one of our own design (Malcolm and Elliott 7).Since a small amount of specimen birefringence is usually present, it is desirable to rotate the specimens round the axis of the polarimeter tube in which they are housed. Initially this was done in steps by hand, and the average of a number of such readings was taken. Later a small motor was fitted to rotate the specimens at a high speed, and only one reading was then required. Careful filtering and drying in filtered air is needed to 167168 INFRA-RED SPECTRA OF POLYPEPTIDE AND PROTEIN FILMS produce clean films. When the number of films required is large (say twenty or more), or when the film thickness is irregular, it is advantageous to immerse the films in a suitable liquid to reduce reflection and distortion of the polarized beam. Edwards' silicone fluid 703 as used for diffusion pumps is suitable for some purposes.Its disadvantage lies in the presence of dissolved air, some of which may form a bubble which, owing to centri- fugal forces, remains on the axis of rotation just in the middk of the field of view. Fre- quent evacuation of the filled tube is needed to remove this air. Styrene is also suitable in some cases. Measurements of the dispersion of optical rotation were made initially with filtered light from a mercury lamp. Later a small monochromator was substituted for the filter, since it is difficult to get sufficient spectral purity and intensity with filters. The films were cast on small squares of thin sheet glass which were annealed after cutting. The weight per unit area was obtained by marking a small central square with a razor blade and subsequently weighing the film contained within this square.Since the total amount of this material was usually only a few milligrams, the principal source of error lies in this determination. It is not necessary to know the specific gravity of the polymer in order to calculate the specific rotation, for the latter quantity may be defined as the rotation produced by 100 g of material in a column of 1 sq. cm cross-section. The accuracy of measurement of optical rotation in films naturally falls short of that which can be obtained with solutions, for with the maximum number of films which can in practice be used, the rotation produced is small. In the work to be described, the angle of rotation varied from 0.04" to about 1.2" ; the polarimeter could be set to about 0-001".RESULTS POLY ALANINE Four polyalanines, copolymers containing different ratios of L- and D-alanine were examined as films. They were cast at room temperature from 5 % solutions in dichlor- acetic acid into an alcohol. To produce the best films, it was found desirable to use isobutyl alcohol for the D-polypeptide, n-propyl alcohol for the next two members of the series and methyl alcohol for the last one. D-residues were present in excess in these polymers, but to facilitate comparison with earlier results 3,6 they are recorded and plotted as if the L-component had been predominant. The films were well washed in ethyl or methyl alcohol and air-dried. Films of suitable thickness were cast on plates of thallium bromo-iodide under the same conditions to provide specimens for infra-red examination.Residue rotationsp corrected for an assumed refractive index of the polymer of 1.5 are shown in fig. 1 plotted against the fraction of L-residues for five different wavelengths. It will be seen that all the observed rotations for polyalanine films are negative, and by comparison with rotations observed for this polymer in solution it might be thought that the polyalanine films were in some form other than an a-helix. This, however, is not the case, for the dispersion is of the " anomalous " type associated with a-helices. As shown by Moffitt and Yang 2 the dispersion of the optical rotation produced by a-helices is of the form and on plotting [RVac](X2 - A@ against 1/(P - A@ a straight line of slope b& is ob- tained.The results here presented do not allow an independent determination of Ao, but when this constant is assumed to have the value 2120A, as found by Moffitt and Yang for poly-y-benzyl-L-glutamate, the curves shown in fig. 2 are obtained. For poly- alanine films whose L/(D + L) composition is 1.0, 0-9 and 0.8 the values of bo are re- spectively - 475, - 500 and - 505 deg. cm2 per decimole. These may be compared with - 630" for poly-y-benzyl-L-glutamate 2 and - 460" for polyleucine with L/(D + L) equal 6 to 0.875. When the optical rotation contains a contribution from an arrangement of a-helices, the plot of [R,,] against L/(D + L) gives a linear part of greater or less length which on extrapolation does not go through the origin.% 6 This is clearly the situation in fig.1. In this graph, for reasons discussed below, the linear relationship appears to be restricted to points for which L/(D + L) is 0-8 or more. Over this range, the linearity of the pIot and the approximate constancy of bo both show that one sense of helix is predominant. The plots extrapolate to intersect the [RvJ axis at points which represent the values of this quantity for a right-handed helix of meso composition. Although such extrapolationFIG. A. ELLIOTT, W. E. HANBY A N D B . R . MALCOLM 169 0.5 0.6 0-7 0.8 0.9 L D*L - 1 .-Optical rotations of solid films of polyalanine plotted against composition, different wavelengths : a, 5780 A ; b, 5461 8, ; c, 4358 A ; d, 4047 A ; e, 3663 FIG. 2. - Dispersion of optical rotation for various solid films.a, poly-L-alanine ; b, poly- alanine L/(D + L) = 0.90 ; c, polyalanine L/(D + L) = 0.80; d, polyalanine L/(D + L) = 0.67 ; e, Bom- byx mori silk (from aqueous solution) ; A Bornbyx mori silk (from dichloracetic acid) ; g , potassium salt of poly-L-glutamic acid ; h, lysozyme. I 08/(X2 -A02 ) 0 5 I 0 for A.170 INFRA-RED SPECTRA OF POLYPEPTIDE AND PROTEIN FILMS is very inexact, i t is interesting to note that the values are all positive, that they increase at first with diminishing wavelength and then become stationary as though passing through a maximum before diminishing again-exactly as has been observed with solutions of a-polypeptides. The value of bo for the meso polymer (subject of course to considerable uncertainty) is - 560".The points corresponding to L/(D + L) equal to 0.67 do not lie on the linear part of the curves, and on plotting [R,a,-](A' - A;) against 1/(P - A;) a graph of nearly zero slope is obtained (fig. 24. This shows that bo is nearly zero, and the anomalous dispersion of the helix is absent. The infra-red spectra of the polyalanine specimens shown in fig. 3 are in complete agreement with the deductions made from the optical rotation measurements. In the spectra of polymers of 1.0, 0.9 and 0.8 L/(D + L) composition, the bands at 893 and 906 cm-1, known to appear in or-poly-L-alanines are well-marked. The 906cm-1 band is particularly important, since it appears to be associated with a crystalline arrangement of or-helices. In the spectrum of polyalanine of 0.67 L/(D + L) composition, this band is absent.There is, however, a strong band at about 966 cm-1, which shows the presence - 9 0 0 I000 900 1000 1100 wave number, cm-1 FIG. 3.-Infra-red spectra of polyalanine films whose optical rotation is given in fig. 1. a, L/(D + L) = 1.0 ; d, L/(D + L) = 0-67. by L/(D + L) = 0.9 ; c, L/(D + L) = 0.8 ; of the /3 form of polyalanine. The carbonyl band in the spectrum of a thin film shows the sharp band at about 1630cm-1 characteristic of /I polypeptides, as well as a broad band centred at about 1655 cm-1. This coincides with the wave-number of the.carbony1 mode of simple synthetic polypeptides in the a-helix form, but as will be shown in the section on the structure of water-soluble silk, it also occurs in the spectrum of an amor- phous, possibly random coiled arrangement of polypeptide chains.Since the dispersion of the optical rotation shows conclusively that or-helices are absent in the film with 0-67 L/(D + L) compoyition, it appears likely that part of this polymer film is randomly coiled. It must be realized that the difference in behaviour between this polyalanine and the other three may not arise solely from the difference in the L/(D + L) ratio. The molecular weight may be different, and in addition it was found necessary to use a quick precipitant (methyl alcohol) to produce suitable films in the one case, whereas with the other three polymers a slower precipitant was used. These results may be compared with some which were obtained earlier for solutions of polyalanine.6 In chloroform containing 1 % dichloracetic acid the rotations are posi- tive for all values of L/(D + L) in excess of 0-5 and the form of the curve shows that helices are present.The fact that we find negative values for [Rva,] in the solid film shows the great effect which the environment has on the value of ao, an effect which has been observed (in a much smaller degree) by Yang and Doty in solutions of poly-y-benzyl-L-glutamate. 1 t is therefore particularly interesting that the values of bo should be similar to those found for other polypeptides. This confirms Moffitt and Yang's contention that bo is much more invariant than ao. The new results also extend the validity of a conclusion drawn from our earlier observations on poly-y-benzyl glutamate that solvent effects on the opticalA .ELLIOTT, W. E . HANBY AND B . R . MALCOLM 171 rotation of an a-helix of meso composition are small. Thus this quantity, given by the intercept on the y-axis of fig. 1 has the value 70" for light of wavelength 5780A, for the solid polymer. The corresponding value (for light of not greatly different wavelength 5893 A) for meso polyalanine in chloroform containing 1 % dichloracetic acid 6 is about 80". From comparison of the X-ray diffraction pattern of solid films of a-poly-L-alanine with the optical transform of the a-helix, it has been shown that the right-handed helix is the dominant one (Elliott and Malcolm 9. lo). It was at first believed that the negative value of bo found in measurements of optical dispersion in some synthetic polypeptides was evidence of this sense in solutions, but it appears doubtful whether this is a valid conc1usion.s However, the fact that similar values for bo are found in polyalanine fi!ms and in a number of solutions of L-synthetic polypeptides must surely mean that the right- handed form is the stable one for the L-enantiomorph. The linear part of the plot in fig.1 shows that for values of L/(D + L) of 0.8 and over, the left-handed form in solid polyalanine is not present to a significant extent. ALKALI SALTS OF POLY-L-GLUTAMIC ACID Films of the sodium and the potassium salts of poly-L-glutamic acid were made by dissolving the poly-acid in aqueous solution containing the stoichiometric amount of alkali, to give solutions of about 13 % concentration (w/v).The solutions, after filtering, were cast on glass plates at about 40°C in a dry air stream. The resulting films are very hygroscopic and it was found desirable to store them in a warm desiccator until a sufficient number had been made. They were then quickly transferred, still warm, to the polari- meter tube which contained phosphorus pentoxide. After measurement, the further manipulation for determining film thickness was done over a warm plate, and the films were well dried at 70°C before weighing in a closed bottle. Fig. 2(g) shows the results obtained for the potassium salt ; for the sodium salt almost identical rcsults were obtained. The small value of the slope of the line is indicative of the absence of any considerable fraction of the polymer in the a-helix configuration.Films of both polymer salts of suitable thickness for infra-red measurement were prepared under similar conditions to those described above, and measured. The carbonyl stretching mode was found at 1658 cm-1 in both cases, hence these results furnish a second example of a polypeptide (in this case of simple composition) in which a carbonyl band near 1660 cm-1 is not associ- ated with the helix form. WATER-SOLUBLE SILK Aqueous dispersions of Bombyx mori silk (made for instance by dissolving the silk in aqueous lithium bromide and dialysing out the salt) may be used to cast films which are soluble in water. Some years ago it was found that the spectra of these films have a carbonyl absorption band at 1660 cm-1 and since this same band is found in the spectra of synthetic a polypeptides it was suggested that water-soluble silk was in the cc form (Ambrose, Bamford, Elliott and Hanby 11).At this time, although the a-helix had been proposed by Pauling and Corey 12 its validity had not been established. The a-helix is now known to be a stable form of the synthetic polypeptides, but although good evidence of this form has been found in water-soluble Antherea nzylitta and Anaphe moloireyi silks (from X-ray diffraction rings in both materials, and from infra-red spectra in the former) it has not been found in films of water-soluble Bombyx silk (Elliott and Malcolm 139 14). The evidence against the a-helix form for this last silk was, however, negative in character, and it appeared that a more positive indication could be obtained from measurements of optical rotation. Dilute aqueous solutions of Boinbyx silk have been shown to be in a random coil form, and the a-helix form has been found in solutions in a suitable mixture of ethylene dichloride and dichloracetic acid.4 Films were cast by placing a few drops of the aqueous dispersion on gIass plates heated to about 70°C on a small rotating table, and drying in a stream of dried, filtered air.Such films are not usually completely soluble in water again after drying, and for good solubility the films should be dried on mercury at 100°C. This is not practicable for films required for optical rotation measurement. The infra-red spectra of silk films dried on a solid substrate at 70°C are almost identical with those of water-soluble films except that a very small shoulder at about 1630cm-1 shows the presence of a small amount of /3 material, which prevents the film from dispersing completely in water.The silk films were much too irregular in surface to use without an immersion medium, and for this purpose styrene was employed. The optical rotation was negative at all wavelengths, and dispersion was found to be of the normal type, which shows either that a-helices are not present or that there are equal numbers of right- and left-handed helices. In172 INFRA-RED SPECTRA OF POLYPEPTIDE AND PROTEIN FILMS view of what is known of the sense of the a-helix in simple polypeptides, this latter possi- bility can be excluded. The most sensitive test for the absence of " form " rotation is furnished by a plot of [Rvac](A* - A;) against 1/(A* - A;).This gives a line of zero slope as shown in fig. 2 when A0 is given the value 2120 A, hence the coefficient bo is also zero. This result is an important one, and shows convincingly that a carbonyl absorption band at 1660 cm-1 does not necessarily indicate an a-helix form, even in a material which can under some circumstances take this configuration. The zero value for bo shows absence of " form " optical rotation. Polypeptides in the extended /3 form may have zero boy as is shown by fig. 2(d) and (f). However, the amount of /3 material present in silk films prepared from aqueous solution is insignificant, and a random arrangement seems likely. In air-dried silk films, an appreciable amount of water is present and it is likely that many peptide groups are hydrogen-bonded to water molecules.However, on heating such silk films to 70°C in a closed cell containing P2O5 for a number of hours much of this water is removed. The infra-red spectrum shows no trace of a free NH band and the carbonyl absorption band remains at 1660 cm-1; hence it must be concluded that the hydrogen-bonding capacity of the silk is satisfied by intra- or inter-chain bonds. This presumably random-coiled form in the solid state must be considered as a stable form of silk, and one which may well be found in other dry proteins. Since it is now established that neither aqueous, dilute dispersions of Bornbyx mori silk nor the films cast under suitable conditions from such solutions contain a-helices, it is most unlikely that aqueous dispersions would have the helical form at some inter- mediate concentration.This conclusion casts considerable doubt on the claim to have established the existence of the a-helix as a major component in such dispersions by methods based on small-angle scattering (Kratky, Sekora and Pilz 15). LYSOZYME It was found possible to cast films of lysozyme from aqueous solution and to measure the dispersion of the optical rotation, and also the infra-red spectrum in the region of the carbonyl stretching mode on specimens prepared under as far as possible identical con- ditions. A rather broad band centred at 1660 cm-1 was observed, with no indication of any peak or shoulder at 1630cm-1. The plot of the optical rotation measurements is shown in fig.2(h), and gives no indication whatever of the kind of dispersion characteristic of a-helices. It is reasonable to conclude that the polypeptide chains in lysozyme films are neither in the extended (p) nor in the a-helix form. Lysozyme is capable of forming single crystals, but it by no means follows that in our films the polypeptide chains are in the same form as in crystals. CONCLUSION Since a polypeptide chain is capable of forming internal and external hydrogen bonds in a number of different folds, it is perhaps not surprising that polypeptide and protein films can be made in which neither the a-helix nor the extended /3 form are detectable, and this is what our experiments show. It is not known whether in such cases there is any degree of order in the films, and it may be that they are simply disordered states. Whatever the state of the polypeptide chains in these films, it is clear that they cannot be distinguished from a-helices by observations of the frequency of the C=O band alone. 1 Doty and Yang, J. Amer. Chem. SOC., 1956, 78,498. 2 Moffitt and Yang, Proc. Nat. Acad. Sci., 1956, 42, 596. 3 Elliott, Hanby and Malcolm, Nature, 1956, 178, 1170. 4 Yang and Doty, J. Amer. Chem. Soc., 1957, 79, 761. 5 Moffitt, Fitts and Kirkwood, Proc. Nat. Acad. Sci., 1957, 43, 723. 6 Downie, Elliott, Hanby and Malcolm, Proc. Roy. SOC. A, 1957, 242, 325. 7 Malcolm and Elliott, J. Sci. Instr., 1957, 34, 48. 8 Elliott, Proc. Roy. SOC. A, 1954, 226,408. 9 Elliott and Malcolm, Nature, 1956, 178, 912. 10 Elliott and Malcolm, to be published. 11 Ambrose, Bamford, Elliott and Hanby, Nature, 1951, 167, 264. 12 Pauling and Corey, Proc. Nat. Acad, Sci., 1951, 37, 235. 13 Elliott and Malcolm, Trans. Faruday SOC., 1956, 52, 528. 14 Elliott and Malcolm, Biochim. Biophys. Acta, 1956, 21,466. 15 Kratky, Sekora and Pilz, 2. Naturforsch., 1954, 9b, 803.