Professional Documents
Culture Documents
1998 Surface Structure of Native Cellulose Microcrystals by AFM
1998 Surface Structure of Native Cellulose Microcrystals by AFM
1998 Surface Structure of Native Cellulose Microcrystals by AFM
Abstract. Atomic force microscopy (AFM) has been used Microdictyon that these two different crystal phases can co-
to study the surface of native cellulose I microcrystals from exist within a single microfibril [4]. One of the unresolved
Valonia ventricosa. High-resolution images show clear struc- problems of structural research into native cellulose I is pre-
tural details of the surface, namely the 0.52 nm repeat along cisely where the triclinic and monoclinic phases exist within
the cellulose chains resulting from the glucose sub-unit and a microfibril [4].
the inter-molecular spacing of ∼ 0.6 nm. Cellulose from Val- We have applied the technique of atomic force mi-
onia exists naturally in both a triclinic (Iα ) and a monoclinic croscopy (AFM) to image the surface of native cellulose I
(Iβ ) crystal form within the same microfibril; the main differ- microcrystals at ultra-high (sub-nanometer) resolution. The
ence between them being a displacement of adjacent chains AFM images under both propanol and water have clearly
by a quarter of the c-axis period to give either a diagonally shown the pitch of 0.52 nm along the molecule due to the
shifted or staggered arrangement of the cellobiose units. The asymmetrical glucose unit and the inter-molecular spacing of
most significant finding in this work is that it has been pos- approximately 0.6 nm.
sible to image the cellobiose repeat along the chain because The topographical differences associated with the screw
of topographic differences associated with the asymmetric symmetry along the cellulose chain have enabled us to iden-
glucose unit, and thus identify triclinic structure on the mi- tify the triclinic phase by direct imaging [5]. To support our
crocrystal surface. Computer modelling has been used to con- interpretation of the images, we have compared them with
struct pseudo-AFM topography images from Connolly sur- computer-generated Connolly surfaces [6, 7] based on crys-
faces of the facets of the two different crystal forms, and the tallographic data from diffraction experiments [4, 8], assum-
triclinic models are in excellent agreement with the real im- ing no molecular relaxation at the surface compared to the
ages obtained by AFM. interior.
Cellulose is a homopolymer of β-[1,4] D-glucose molecules Valonia ventricosa was harvested from a seabed in Florida
linked in a linear chain, with alternating sub-units in crys- Keys, and the microcrystals were prepared by using the
talline cellulose rotated through 180◦ . Native cellulose I method from Revol et al. [9]. Cellulose fragments were dis-
microfibrils are highly ordered crystals and evidence from integrated in a homogeniser and sulphuric acid was added
13
C NMR spectroscopy and electron diffraction suggests that slowly to improve the dispersion. The mixture was then heat-
they consist of both triclinic (Iα ) and monoclinic (Iβ ) allo- ed to 70 ◦ C with strong stirring for 30 minutes. The resultant
morphs [1–4]. In both allomorphs, the molecules have very dispersion was washed with distilled water and finally a drop
close to strict two-fold screw symmetry along the chains, of chloroform was added before storage as a protectant.
which means that the true repeating unit is cellobiose, com- Samples of Valonia for atomic force microscopy were
prising two glucose units. The difference between them is prepared by depositing a 5 µl drop of a diluted sample of
a shift of adjacent hydrogen bonded sheets along the chain the cellulose (approximately 20 µg/ml) onto freshly cleaved
axis, resulting in either a stagger or a diagonal shift of the mica. For imaging under water, APTES-mica was used to im-
cellobiose unit by a quarter of the c-axis period. prove the binding of the cellulose to the mica substrate. The
The cellulose used in this work is from the green alga Val- APTES-mica was prepared by exposing freshly cleaved mica
onia ventricosa, consisting of Iα and Iβ phases roughly in the to an atmosphere of aminopropyltriethoxysilane for several
ratio 65 : 35 [2]. It is known from studies of cellulose I from hours [10].
S560
1 µm
triclinic unit cells were constructed according to the parame-
ters of Sugiyama et al. with atomic co-ordinates from electron
diffraction [4], which were further characterised by molecu-
lar dynamics simulations [8]. The crystal structures were built
Fig. 1. AFM topography image taken in contact mode under propanol show-
in Cerius2 TM (Molecular Simulations Incorporated). Crystal ing a typical field of Valonia cellulose microcrystals
surfaces were then cleaved from these bulk structures, and
the Connolly surface generation algorithm in the Cerius2 TM
program used to produce a surface [6, 7]. The size of the
probe rolling over the surface in the figures presented here
was 0.2 nm, comparable with the size of a water molecule.
2 nm
Larger probe sizes simply smoothed the surface and revealed
less detail, but did not obscure the key structural features of
the surface. The size of the probe is considerably smaller than
the 20–50 nm average radius of a typical AFM tip (Digital In-
struments specification), but the high resolution of the images
obtained is very likely to result from the presence of much
smaller micro-asperities at the end of the tip.
The generated surfaces were exported to custom-written
software to convert the surface co-ordinates into a pseudo-
AFM topography image.
at 1.09 nm, in close agreement with the cellobiose repeat interval. No ordi-
nate scale is shown since the line profile is taken across a deflection image,
and the grey-scale has no direct relationship to height. c Fourier transform
along the line profile highlighting the spectral peak at 1.09 nm
63o 67o
image between the two marker arrows, averaged across the
width of the chain. It should be noted that this is a line pro-
file from a deflection image, so the ordinate of the line profile 0.53 nm 0.62 nm
does not correspond to height changes and is presented in ar- (d010) (d100)
bitrary units. The prominent peaks are separated by 1.09 nm,
Monoclinic (110) Monoclinic (110)
as is clearly shown in the Fourier transform of the line pro-
file in Fig. 3c. This pattern can easily be seen along the chains
adjacent to the one profiled. A similar analysis from a line
1.04 nm
necessary to make the comparisons in reciprocal space [19]. bands. The line profile shown in Fig. 5b, averaged across the
The similarity of the triclinic Connolly surfaces shown in width of the dotted white box, again clearly shows the repeat
Fig. 4a,b with several areas of the image shown in Fig. 3a will of about 1 nm between the most prominent peaks. Most of
be apparent immediately. This pattern is not visible over the these peaks are split into two, for example on the right-hand
entire image, possibly because of some disruption of the cel- extreme of the line profile. These larger and smaller inter-
lulose surface. It is also difficult to achieve the highest quality vals appear prominently on the Fourier transform of the line
contrast over the entire image at this resolution because of profile.
noise and scanning influences. As referred to in the introduc- This again shows the effect of the two-fold screw sym-
tion, cellulose from this source is likely to show other areas metry along the chain. It is interesting that in Fig. 2 there is
of the surface with monoclinic character (although the tri- apparently no increased significance given to the cellobiose
clinic phase is dominant) and this is the subject of on-going repeat. We have obtained many images of the cellulose sur-
investigations. Although the AFM image shown is a deflec- face, some of which emphasise the cellobiose interval whilst
tion image, the contrast will primarily be due to topographic others do not seem to stress the difference. The interaction
features. Other influences are expected to be present such as between the tip and the surface, as well as the scanning and
elastic deformation and frictional forces [20, 21], but these do feedback parameters are certainly very important in this re-
not detract from the fact that alternate sub-units are inverted gard. The change of environment from propanol to water may
along the cellulose chain, which manifests at the crystal sur- also be important; either because the chemical groups ex-
face as a topographic change. posed to the surface are surrounded by a different medium,
Figure 5 supports this hypothesis further. This deflection or because the tip–sample interaction is modified. Further ex-
image of the Valonia surface was taken under water, with the periments are in progress to assess this.
chain axis parallel to the dotted white box. A repeating feature
can be seen with some qualitative indication that the bright,
wide band is split in two, such that there is a “pairing” of sub- 3 Conclusions
References
1. R.H. Atalla, D.L. VanderHart: Science 223, 283 (1984)
2. D.L. VanderHart, R.H. Atalla: Macromolecules 17, 1465 (1984)
1.13 nm c 3. J. Sugiyama, T. Okano, H. Yamamoto, F. Horii: Macromolecules 23,
0.56 nm 3196 (1990)
4. J. Sugiyama, R. Vuong, H. Chanzy: Macromolecules 24, 4168 (1991)
5. A.A. Baker, W. Helbert, J. Sugiyama, M.J. Miles: J. Struct. Biol.
1 nm 119(2), 129 (1997).
b
6. M.L. Connolly: Science 221, 709 (1983)
Fig. 5. a A deflection image of the cellulose surface taken under water. 7. M.L. Connolly: J. Appl. Crystallogr. 16, 548 (1983)
The chain axis is parallel to the dotted white box which indicates where 8. A.P. Heiner, J. Sugiyama, O. Teleman: Carbohydr. Res. 273, 207
the line profile was taken from in b. There is a qualitative indication that (1995)
the bright, wide bands are split into smaller sub-bands. In b, this band- 9. J.-F. Revol, H. Bradford, J. Giasson, R.H. Marchessault, D.G. Gray:
ing can be clearly seen along the line profile. The major repeat of 1.13 nm Int. J. Biol. Macromol. 14, 170 (1992)
(black arrows) is sub-divided by a smaller repeat of 0.56 nm (grey paired 10. Y.L. Lyubchenko, B.L. Jacobs, S.M. Lindsay: Nucl. Acids Res. 20,
arrows). This observation supports the idea that alternate glucose units are 3983 (1992)
exposing a more prominent chemical group on the surface and making the 11. C.A.J. Putman, K. van der Werf, B.G. de Grooth, N.F. van Hulst,
cellobiose repeat more significant. c shows the Fourier transform along the J. Greve, P.K. Hansma: Proc. SPIE 1639, 198 (1992)
line profile in b 12. J.-F. Revol: Carbohydr. Polym. 2, 123 (1982)
S563
13. J. Sugiyama, H. Harada, Y. Fujiyoshi, N. Uyeda: Planta 166, 161 18. E.M. Debzi, H. Chanzy, J. Sugiyama, P. Tekely, G. Excoffier: Macro-
(1985) molecules 24, 6816 (1991)
14. S.J. Hanley, J. Giasson, J.-F. Revol, D.G. Gray: Polymer 33, 4639 19. L. Kuutti, J. Peltonen, J. Pere, O. Teleman: J. Microsc. 178, 1 (1994)
(1992) 20. R.M. Overney, E. Meyer, J. Frommer, D. Brodbeck, R. Lüthi, L.
15. D.J. Keller, F.S. Franke: Surf. Sci. 294, 409 (1993) Howald, H.-J. Güntherodt, M. Fujihira, H. Takano, Y. Gotoh: Nature
16. P. Mulvaney, M. Giersig: J. Chem. Soc., Faraday Trans. 92, 3137 359, 133 (1992)
(1996) 21. J.H. Hoh, C.-A. Schoenenberger: J. Cell Sci. 107, 1105 (1994)
17. H. Yamamoto, F. Horii, H. Odani: Macromolecules 22, 4130 (1989)