Peptides and amino acid structures

Date:
1941-1959
Reference:
PP/CRI/H/2/35
Catalogue details

Licence: In copyright

Credit: Peptides and amino acid structures. Source: Wellcome Collection.

    4/133
    Henri A. Levy and Robert E. Corey Vol. 63 value within the limits of error of the determina tion. A knowledge of corresponding atomic separations in the molecule of alanine, in which a hydrogen atom is replaced by a CH 3 side chain, seemed a desirable extension of our information regarding these fundamental units of protein structure. For this reason the present deter mination of the crystal structure of dl- alanine was undertaken. In a preliminary paper upon the crystal struc tures of some amino acids and related compounds, Bernal 4 reported that the unit cell of ¿/-alanine is orthorhombic with a 0 = 6.0 Â., b 0 = 12.0 Â., Co = 5.8 Â., containing four molecules of CH 3 NH 2 - CHCOOH. He found the space group to be Cl, — Pbn- Th e crystals are optically negative with a parallel to the long axis, b. The general features of a suggested arrangement of the mole cules within the unit cell were depicted, but no X-ray intensity data were given and no attempt was made to determine the positions of the atoms. Preparation of Crystals.—The dl- alanine used was ob tained from Amino Acid Manufactures, University of California at Los Angeles. The crystals, which were grown by slow evaporation of aqueous solutions, were small, well-formed needles elongated along the c axis, and charac terized by the absence of any pronounced cleavage. For photographs involving rotations about this axis specimens For rotations about the a axis it became necessary to mount somewhat larger crystals (m. 0.5 mm. diameter) horizontally on a firm glass fiber by means of shellac which was then thoroughly hardened by baking at about 70°. The orientation of the crystal was then adjusted by means of an optical goniometer and checked by X-ray pho tography, after which the ends of the needle were carefully dissolved away leaving a tiny portion in the form of a verti cal cylinder of approximately uniform cross section which was used for the collection of X-ray intensity data. The Unit Cell and Space Group. —Weissenberg photographs taken about the a and c crystallo- graphic axes showed the unit cell of ¿/-alanine to be orthorhombic. Measurements of reflections from (200), (210), (310), (320), (002), and (011) observed on rotation photographs were used to ob tain the dimensions <h = 12.04 Â., b<¡ = 6.04 Â., Co = 5.81 À., all =fc 0.01 À., based upon the wave length 1.539 Â. for the Cu Ka doublet. The den sity of the crystals, determined by the flotation method, was found to be 1.40, so that the unit of structure contains four molecules (calculated 4.00). The presence of all orders of (hkl) indi cates a primitive lattice. Systematic absence of all {hOt) reflections having h odd and of all (Qkl) reflections having k + I odd establishes the space group as DU ~ Pnam or Cl, — Pna. Of the three sets of four equivalent positions provided by — Pnam, two require that the atoms lie at symmetry centers 000, 00J, |§0, J§§, an d 0|0, Oli, è00, i0|, and the third requires that the atoms lie in mirror planes perpendicular to the c axis, at xyi; xyí; i ~ x, i + y, î; h + x, h - y, î- Since it is certain that an alanine molecule has neither a center nor a plane of symmetry, the atoms of the four alanine molecules cannot lie in these positions. Restrictions of this sort are not imposed by the symmetry elements of C\, — Pna, in which all atoms occupy the general posi tions xyz; X, y, J + z; è - x, è + y, è + z; | + X, J — y, z. Pna is thus the space group which is to be considered as a basis for this structure. The symmetry elements present in the space group Pna are (1) two-fold screw axes at xy = 00, §0, 0.¿, and JJ, (2) n glide planes at x — \ and f, and (3) a glide planes at y = j and }. Axial ratios of ¿/-alanine calculated from our X-ray measurements are a : b : c = 1.993:1:0.962. Those based upon optical measurements reported by Groth 5 are 0.9971:1:0.4949, which, upon halving the length of the b axis, become 1.9942: 1:0.9898. Difficulty in obtaining satisfactory optical signals from the terminal faces of the needle-like crystals is doubtless responsible for the discrepancy of about 3% in the c:b ratio and also for the erroneous assignment of alanine to the point group D lh . Upon interchanging the a and b axes both the size of the unit and the space group are seen to be in accord with the results previously reported by Bernal. 4 Determination of the Structure Disregarding the hydrogen atoms, which con tribute too little to the X-ray scattering to per mit the determination of their positions, a solu tion of the structure requires the assignment of values to three parameters, x, y, and z, for each of two oxygen atoms, three carbon atoms, and one nitrogen atom, with the exception that the sym metry of the space group Pna allows one z pa rameter to be arbitrarily assigned. Preliminary examination of the X-ray diffraction data shows that reflections from the planes (210) and (400) are strong, while that from (120) is weak. These data strongly suggest a high concentration of scattering matter in the neighborhood of x = Aug., 1941 Crystal Structure of dl- Alanine 2097 Vs, y = l /i, with an approximately symmetric distribution in the y direction about y = í /t, The optical data 4 suggest compression of the distri bution in the x direction, with greater extension in the y and z directions. Although these clues to the positions of the atoms together with knowledge of the size and shape of the glycine molecule 2 might be expected to provide a valuable guide in the determination of the atomic arrangement in crystals of dl- alanine, the presence of the methyl group intro duces so many complications into the structure that any attempt at its solution by trial and error methods was soon found to be impracticable. X-Ray intensity data from a complete series of Weissenberg photographs taken about the a and c axes using Cu K a radiation were therefore com piled and used in the calculation of sections through the Patterson diagram 8 as suggested by Harker. 7,2 The data employed in the evaluation of the various F* series may be obtained from Tables I, II,. Ill, and IV. Estimations of the relative intensities of re flections were made by comparing densities of spots on several negatives which differed only by known exposure ratios. In the case of the (hh0) spectra, photographs yielding intensity estima tions believed to be especially reliable were ob tained by a multiple film technique similar to that employed by deLange, Robertson, and Wood ward. 8 Five Agfa non-screen X-ray films were placed one behind another in a Weissenberg camera and exposed simultaneously. The con stant ratio of exposures of corresponding spots on adjacent films was found to be approximately 2.8 for Cu Ka radiation. Values of F 2 were ob tained from the estimated intensities by dividing them by the Lorentz and polarization factor for equi-inclination Weissenberg data, (1 + cos 2 2i?)/sin 2# sin ip, in which ê is the Bragg angle and <p the azimuthal angle of the reflection. The three-dimensional Patterson function P(X, Y, Z) = 2ZS I Fu, 1 2 cos 2 T(hX + kY + IZ) origin which correspond to vector distances be tween atoms in the crystal. (The symbols X, Y, and Z which represent the arguments of the Pat terson function should not be confused with x, y, and z, the parameters of position in the unit cell.) The presence of two-fold screw axes per pendicular to the xy plane suggests that the dia gram most useful for the determination of atomic positions should be the section Z = | of the Pat terson diagram, given by the series P(X, Y, è) = 222 I F m I » cos 2 r (hX + kY + U) which may be rewritten in the form P{X, Y, y = 22 Cu COS 2 t (hX + kY) in which Cm = Z(-1)H,Fm |» In the present case the maxima of this function will arise from the following types of interatomic distances in the crystal. (1) Two atoms related by a two-fold screw axis will have a separation of one-half unit in the z di rection and will give rise to a peak in P{X, Y, |). As has been noted, 7 ' 2 one-half of the vector from the origin to such a maximum may therefore be interpreted directly as the vector displacement of an atom from one of the screw axes of the crys tal. If, then, a plot of P(X, Y, |) reduced to half scale is placed upon a projection of the unit cell so that the origin coincides with the positions of the screw axes, a maximum of this type might be expected to represent an atomic position. Since the function is unable to distinguish be tween the two non-equivalent sets of screw axes of the space group Pna, the number of maxima of this sort will be in general twice the number of atoms producing them. (2) Any pair of atoms having a separation of about one-half unit in the z direction will give rise to a maximum in the function P(X, Y, |). (3) Two atoms lying in or near the same xy plane will give rise not only to their respective peaks of type 1, but to an additional maximum half way between these two. 2 Fortunately the atomic distribution in crystals of ¿/-alanine is such that no maxima of this type are observed. (4) Two atoms related by the n glide plane will have a separation of one-half unit in both the y and z directions and will therefore give rise to a maximum in the function P(X, Y, i) lying along the line Y = J. One-half the dis tance from the Y axis to a maximum along this line may thus represent the distance of an atom from the n glide plane. The general features of the diagram, shown in Fig. 1, suggest a concentration of atoms near x = l, y = i in the unit cell, in agreement witli