Organization of microfibrils in keratin fibers studied by X-ray scattering: Modelling using the paracrystal concept

https://doi.org/10.1016/S0167-4838(98)00216-7Get rights and content

Abstract

Low-angle X-ray scattering patterns of hard alpha-keratin fibers have been studied for more than 50 years but a completely convincing modelling has never been presented. The models which have been proposed so far are specific to the sample and cannot be adapted to others, mainly because they do not use a parametric analytical expression of the distribution function describing the relative positions of the microfibrils. Our new approach is based on a paracrystal distribution function. In addition, a huge background originating from a non-ordered matrix is taken into account. Various hard alpha-keratins from different origins have been studied using our approach. From the rather good modellings obtained, it appears that the diameter of the microfibril is not origin dependent (7.4 nm) whereas the distances between microfibrils and their electron density profiles are. Hair microfibrils can be reasonably approximated by a solid cylinder but a core and an outer ring are necessary for porcupine. Our method is of course not limited to keratin microfibrils; it can be used for modelling equatorial X-ray scattering profiles of all types of hexagonal fibrillar assemblies, which are in fact widely found in biological tissues.

Introduction

Fibrous tissues are widespread features in biology, as for instance collagen, cellulose or intermediate filaments. As illustrated by the example of hard alpha-keratin (the major component of many tissues such as hair, nails and quills), the fibrous structure is often simultaneously present at different scales, from the molecular level (alpha-helical coiled coil) up to the macroscopic one (hair). The basic unit of fibrous tissues is the fibril (or microfibril according to the terminology of each compound) with diameters lying between 5 nm and 100 nm. The fibrillar organization confers outstanding mechanical properties to the tissues. It is, therefore, important to understand the structure of natural fibers, firstly for basic biological reasons and possible applications in pharmacology and medicine, and secondly for helping the design and development of similar synthetic materials.

Electron microscopy and X-ray scattering are the two major techniques for investigating fibrillar organization. Electron microscopy provides a lot of information but is limited to the surface of the sample and is subject to artefacts due to the chemical agents added for sample preparation. In addition, biological materials are generally strongly damaged by electron beams and it is normally not possible to study a unique sample while varying an external parameter on the microscope stage. These limitations do not exist for X-ray scattering and radiation damage is reduced, but data interpretation is not as straightforward because information is obtained in the reciprocal space and also contains extra scattering components from non fibrillar parts superimposed on the fibrillar signal.

Although the model presented in this paper is applicable to any tissue made of an hexagonal two-dimensional lattice of cylinder-shaped units, the present X-ray scattering study is limited to hard alpha-keratin microfibrils organization. This kind of keratin has been and still is extensively studied because it gives rise to scattering patterns made up of many intense and well localized signals due to regular organizations at molecular and supramolecular scales. It is therefore a good candidate for testing a modelling method. In hard alpha-keratin fibers, the 45 nm long molecules are characterized by an alternation of helical parts and non helical segments. These molecules are assembled both longitudinally and laterally forming a very complex supercoil structure called a microfibril [1]. The exact organization of the keratin chains within a microfibril, in particular the putative existence of an intermediate organization level between coiled coils and microfibrils, giving rise to the so called protofibrils, is not yet definitely established.

Bundles of microfibrils embedded in a high sulphur protein matrix, called intermicrofibrillar matrix, form a macrofibril. The macrofibrils are themselves separated by an intermacrofibrillar matrix.

The dense lateral packing of microfibrils produces the intense and rather broad scattering signals located on the equator (perpendicular to the microfibril axes) on X-ray diffraction patterns. A typical equatorial intensity profile I(S), S being the scattering vector (S=2 sin θ/λ, where 2θ is the scattering angle and λ the wavelength), which has been extracted from a two-dimensional pattern of human hair in a glass capillary (Fig. 1A) is shown on Fig. 1B: three broad maxima are superimposed on a huge small-angle signal decreasing monotonically for increasing S values (the parasitic scattering from air or the glass capillary is much weaker than this huge background signal). The peaks are located at S=0.12, 0.22 and 0.34 nm−1 (respectively corresponding to the distances 9.1, 4.5 and 2.95 nm). The thinner peak located at S=0.25 nm−1 has been attributed to crystallized lipid granules which are out of the scope of our present study [2].

Up to now very few interpretations of these equatorial scattering features of hard alpha-keratin have been published. More than thirty years ago, Fraser et al. showed that they are related to the lateral packing of the microfibrils and to their radial electron density profile. They developed the most complete model so far, approximating microfibrils by cylinders located at the nodes of a distorted two-dimensional hexagonal lattice, as seen on microscopy images [3]. The analysis proved that the 0.11 nm−1 maximum could be attributed to the lateral packing between microfibrils whereas the other two were mainly due to their radial electron density profile. This model led to a satisfactory agreement for the positions of the scattering peaks, but to a rather poor one for the intensity profile. Nevertheless it still holds as the best one since no major improvement was afterwards introduced. The most controversial point concerns the localization of the keratin molecules within the microfibril and various models have been proposed. Recently, Parry pointed out that the position of the scattering peaks could be accounted for by various solutions differing in the relative radial electron densities of the outer and inner parts of the microfibril, but the peak intensities were not discussed [4]. The huge small-angle signal has also never been interpreted.

The present study is an attempt to understand the origin of the problems encountered in the interpretation of the equatorial small-angle patterns of hard alpha-keratin and to develop a general model which could lead to a good agreement, for both peak positions and intensities, using a reduced number of parameters. Contrary to the models used so far, which are specific of the studied sample and do not take into account the small-angle background, we propose here a new approach based on a paracrystal distribution function; it also includes an additional scattering contribution due to a non-ordered matrix. This model is then satisfactorily compared to experimental scattering data from several hard alpha-keratins (human and horse hair, porcupine quill).

Section snippets

Samples

Human hairs and horse hairs (from the tail) were analyzed without any special chemical preparation except a short cleaning in water. For the scattering experiments, about 50 to 80 human hairs or 15 to 30 horse hairs were inserted into 1 or 1.5 mm diameter and 0.01 mm thick glass capillaries. Porcupine quills (Atherurus africanus), which are known to display very rich diffraction patterns, were analyzed close to their tip. Temperature analyses were carried out using a Mettler heating stage in

General expression of the scattered intensity

The scattered intensity is proportional to the square modulus of the scattered amplitude (i.e. the Fourier transform of the electron density). The scattering vector range which has been recorded corresponds to the microfibrils diameters. Consequently, we have to model the electron density at the scale of a set of microfibrils embedded within the intermicrofibrillar matrix, only considering a two-dimensional section since the structure will be supposed uniform along the fibril axis. In the

Application to various keratin tissues

The sensitivity and accuracy of the modellings have been estimated following the effect of slight deviations of 2r, 〈a〉 and Δ. The profiles are significantly different for 2r and 〈a〉 shifts larger than 0.1 nm and for Δ shifts larger than 2%. These values apply to all the models presented below.

The paracrystal description

Up to now, two approaches have been used in literature to model the interference function of microfibrils. The first one consists in calculating the Fourier transform of a distribution function estimated from an electron microscopy image. It has been applied to collagen fibers [8] and hard alpha-keratin [3], the experimental interference functions looking in both cases like theoretical interference functions from a paracrystal (Fig. 2). This purely experimental technique is quite correct but

Conclusion

We have presented a new modelling of the low-angle X-ray equatorial scattering intensity profiles of hard alpha-keratin fibers. Three terms are taken into account: the electron density profile of the microfibril, the distribution function of the microfibrils and a background. For the distribution function we have chosen the paracrystal description and adapted it to the case of an hexagonal lattice in order to get a simple analytical function depending only on two parameters: the lattice

Acknowledgements

We would like to thank Dr. Michel Tranier from the Muséum d’Histoire Naturelle, Paris, for providing us with the porcupine quill samples.

References (19)

  • D.A.D. Parry

    Int. J. Biol. Macromol.

    (1996)
  • E.F. Eikenberry et al.

    Int. J. Biol. Macromol.

    (1982)
  • V.J. James et al.

    Int. J. Biol. Macromol.

    (1995)
  • K.E. Wilk et al.

    Biochim. Biophys. Acta

    (1995)
  • R.E. Burge

    J. Mol. Biol.

    (1963)
  • L.N. Jones et al.

    Biophysical Chemistry

    (1997)
  • D.A.D. Parry, P.M. Steinert, Intermediate Filament Structure, Springer-Verlag,...
  • R.D.B. Fraser et al.

    J. Mol. Biol.

    (1963)
  • R.D.B. Fraser et al.

    J. Mol. Biol.

    (1964)
There are more references available in the full text version of this article.

Cited by (54)

  • Keratin films for ocular surface reconstruction

    2011, Biomaterials
    Citation Excerpt :

    X-ray studies did not reveal reflexes in any samples. Thus, we assume that none of the keratin film variations results in a structured arrangement of keratin filaments, as can be found in hair in its native state [36–38]. Similar results for films generated from wool and human hair keratin have been described by Aluigi et al. and Fujii et al. [39,40].

View all citing articles on Scopus
View full text