Single crystals of -chitin J. E. Persson*, A. Domardt and H. Chanzy:~ Centre de Recherches sur les Macromolbcules Vbgbtales, CNRS, BP 53X, 38041 Grenoble cedex, France (affiliated with the Joseph Fourier University of Grenoble)
(Received 31 October 1991; revised 20 February 1992)) Single crystals of ~t-chitin were grown by the addition of precipitants to dilute solutions of low molecular weight chitin fractions dissolved in aqueous LiSCN. At temperatures around 200°C, bundles of thin needle-shaped crystals were obtained. Each of these needles was an or-chitin single crystal, characterized by a spot electron diffraction pattern which could be indexed along the hkO reciprocal net corresponding to the Minke and Blackwell unit cell [a = 0.474 nm, b = 1.88 nm, c (fibre axis) = 1.032 nm, space group P212121]. In a crystal, the a* parameter was along the crystal axis and the b* perpendicular to it. Keywords: ~-Chitin; crystalstructure;i.r. spectroscopy
Introduction During the past 60 years, several studies have focused on the crystallography of ~-chitin 1-6. Among these reports, the most recent determination has been given by Minke and BlackwelP who have presented the latest molecular structure and atomic coordinates of this chitin allomorph. Their model which is refined in the space group P21212~, involves two antiparallel chitin chains per unit cell and one N-acetyl D-glucosamine (2-deoxy2-acetamido-fl, D-glucose) as the independent residue. One of the important features of the Minke/Blackwell model is that it contains a statistical distribution of two rotational conformations for the C H 2 O H moieties, each of them having an occupancy of one-half. This distribution has the advantage of allowing this group to be hydrogen bonded to the remainder of the structure. Such hydrogen bonding is required to confirm the results from infrared analysis, which indicate an absence of free O H groups in a-chitin 7. According to Minke and Blackwell 1, the perfect statistics of the two conformations at C H 2 O H is established only in large a-chitin crystals. In small ones, defects occurring especially at the crystal edges should, in particular, account for the 001 reflection which is forbidden in the P2~2~2: space group. In an electron diffraction study performed on the large chitin crystals 8 found in the grasping spines of Sagitta, Atkins et al. 9 were able to record a series of spot patterns corresponding to the 0kl and h01 sections of the reciprocal space of ~t-chitin. Quite remarkably, these diagrams displayed both 001 (with 1 odd) and 0k0 (with k odd) reflections, in contradiction to the expected diagrams where these reflections should be absent if the proposed P2~2a21 space group was correct. There are therefore conflicting indications concerning the crystal *Present address: Department of PolymerTechnology,Royal Institute of Technology,S-100 44 Stockholm, Sweden. tPresent address: Laboratoire d'Etudes des Mat6riaux Plastiques et des Biomat6riaux, URA CNRS No. 507, Universit6 Claude Bernard, 43, Bvd du 11 Novembre 1918, 69622 Villeurbanne cedex, France. :~To whom correspondenceshould be addressed. 0141-8130/92/040221-04 © 1992 Butterworth-HeinemannLimited
structure of a-chitin. Further clarification requires the preparation of a fair number of crystals of high perfection. From these, one should collect a series of meaningful data by infrared and ~3C C P / M A S n.m.r, spectroscopy, together with X-ray and electron diffraction. It is from the convergence of results from all these techniques, that the final description of the structure of a-chitin will emerge. As quantities of Sagitta grasping spines are not easy to collect, the preparation of single crystals of a-chitin was attempted. For this, a crystallization scheme was devised which follows the general rules devised in our laboratory for obtaining single crystals of polysaccharides ~°. The method involves the preparation of dilute solutions of low molecular weight chitin fractions, followed by their recrystallization at high temperature by the addition of a precipitant. The crystals obtained in such a way, together with their diffraction diagrams are described below.
Experimental Preparation o f chitin o f low degree o f polymerization ( D P )
Chitin is difficult to solubilize and its separation into low D P fractions is therefore hard to obtain. In this work, these fractions were prepared by N-acetylation of low D P chitosan (prepared as in the method described in previous papers 11,1z), using the selective re-acetylation method of Hirano et al. 13. For this, 70 mg of low D P chitosan (DPv = 3511) was dissolved in 1.4 ml of 10% (w/v) acetic acid. A volume of 7 ml of methanol was then added followed by 0.11 ml of acetic anhydride. The mixture was left at room temperature overnight during which a chitin gel formed. This gel was slurried in 100 ml of isopropanol for 24 h. The chitin precipitate was then isolated by centrifugation, re-dispersed in water and freeze-dried. This product, checked by infrared spectroscopy, was identical to an c~-chitin standard extracted from crab tendon. It was soluble in dimethyl acetamide saturated with LiC1 but not in water. In addition, its proton n.m.r, spectrum in D 2 0 / L i S C N gave the expected ratio of three protons at CH 3 to one at NH.
Int. J. Biol. Macromol., 1992, Vol. 14, August
221
Single crystals of o-chitin: J. E. Persson et al. Preparation of the e-chitin single crystals
Infrared spectroscopy
Several systems of solvent/precipitant were tested for the preparation of chitin crystals. Solutions of chitin were prepared in hexafluoroisopropanol, dimethyl acetamide saturated with LiC1, aqueous LiSCN and aqueous LiI. These solutions were recrystallized by adding precipitants such as water, alcohols, pyridine, acetone, dibenzyl ether, etc. at temperatures ranging from ambient to 250°C. The best crystals were prepared by adding aqueous poly(ethylene glycol) ( P E G 200) to a solution of low D P chitin in aqueous LiSCN, at temperatures between 200 and 220°C. Two samples of LiSCN were used: a batch from Fluka was used without purification; another batch from Aldrich did not give as good results as the Fluka LiSCN despite extensive dissolution, filtration and recrysallization. As LiSCN decomposes at high temperature to yield a sulphur-based precipitate which will contaminate the chitin crystals, a limited time is available for recrystallizing chitin. Our best crystallization recipe was as follows: 5 mg of low D P chitin was dissolved in 10 ml of an aqueous solution of LiSCN. This solution was prepared by adding water to LiSCN until its refractive index was 1.428 at 25°C. Then 0.5 ml of the chitin solution was poured into a thin-walled glass ampoule which was sealed and placed inside a small poly(tetrafluoroethylene) ( P T F E ) container, fitted with a P T F E screw cap. A small glass ball whose purpose was to break the ampoule was also put into the container. A measure of 2.5 ml of dilute aqueous P E G 200 (80% water v/v) was poured into the P T F E container which was screw-capped and inserted inside a steel bomb to which a small amount of dilute aqueous P E G was added to produce a counter pressure. The bomb was tightly sealed and immersed in an oil bath kept at 240°C. After 5 min at this temperature, the bomb was shaken vigorously to break the glass ampoule, which caused the chitin solution to be mixed with its precipitant, thus initiating the crystallization. The bomb was then immersed and put into a second oil bath heated to 220°C. After 30 min, the bomb was removed from the oil bath and cooled under tap water. It was opened and the suspension of chitin crystals was removed from the P T F E container. The crystals were washed by successive centrifugation, first in aqueous P E G 200 (80% water), then in water and finally in methanol. The crystals were then stored in methanol.
Infrared spectra were recorded in transmission mode with a resolution of 2 c m - 1. These spectra were recorded on a Perkin-Elmer 1720X FT-IR spectrometer using a microfocus accessory with a micro sample holder disc having a 500 #m aperture. For this, drops of chitin crystal suspensions were evaporated inside a poly(ethylene) cap to yield a thin film which could be handled at the surface of water and thus mounted across the hole of the sample holder.
Results and discussion A typical preparation of o-chitin crystals is shown in Figure 1. This electron micrograph displays a series of bundles each of them consisting of several dozens of needle-like crystals. Within a bundle, the crystals are tightly packed in the middle of the bundle but fan out in a typical sheaf morphology at both bundle tips. The bundles occur either isolated or associated with other bundles. In the latter case, the bundles radiate from a common centre. As seen in Figure 1, a typical bundle has a length of a few #m and a width of about 100 nm in its centre and around 500 nm at each of its ends. The crystals constituting the bundle have pointed tips. They run the full length of the bundle, but have widths that do not exceed 10-20 nm. Their thickness, deduced from images shadowed at an angle, is also in the order of 10 nm. When tested by selected area electron diffraction, a crystal bundle yields a poorly resolved electron diffraction pattern. This is illustrated in Figure 2 where the inserted diffractogram corresponds to the area circled in the image. Despite its poor resolution and substantial arcing, this diagram can be identified as corresponding to the hk0 diffraction arcs from o-chitin indexed along the unit cell defined by Minke and Blackwellt. The bundle axis is aligned along the a* axis of o-chitin and the b* axis is perpendicular to it. The arciffg of the pattern corresponds to the curvature taken by the crystals which fan out within the bundle. Patterns corresponding to one a-chitin crystal can be obtained by using the microdiffraction technique of Riecke and Ruska 15. This is exemplified in Figure 3 which also shows a schematic diagram and proper indexation. The pattern consists of spots which are mirrored around
Electron microscopy and electron diffraction analysis Drops of chitin crystal suspensions were deposited on carbon coated grids and allowed to dry. All electron microscopy and electron diffraction analysis was achieved with a Philips EM 400T electron microscope operated at 120 kV for diffraction purposes and 80 kV for imaging. Electron diffractograms were recorded on Mitsubishi electron image ( M E M ) films and images were produced with Kodak 4489 emulsion. The diffraction diagrams were obtained either with the conventional selected area technique of le Poole 14 or the microdiffraction method of Riecke and Ruska 15. The former method gives information on at least 1/xm 2 of the specimen whereas the latter is appropriate to obtain diffraction patterns on areas as small as a few tens of nm. Some of the electron diffractograms were calibrated by depositing the chitin crystals on carboncoated grids on which a thin layer of gold had been deposited by evaporation.
222
Int. J. Biol. Macromol., 1992, Vol. 14, August
Figure 1 Electron micrograph of a preparation of ~-chitin crystals prepared at 220°C by adding a precipitant (aqueous poly(ethylene glycol), PEG 200) to a dilute solution, of low molecular weight chitin in aqueous LiSCN
Single crystals of a-chitin." J. E. Persson et al. two orthogonal axes a* and b*, a* corresponding to the crystal axis and b* is perpendicular to it. The pattern in Fioure 3 has a resolution approaching 0.2 nm. It displays nine independent spots which can be seen in each quadrant. Quite interestingly, only even reflections (200 along a* and 020 and 080 along b* ) are seen on the two main axes of the diagram. Thus, this diffractogram is consistent with a pgg two-dimensional space group, which itself would meet the P212~2~ requirements provided that one could demonstrate that (a) c* is perpendicular to the a'b* plane, (b) Fhk I = Fffhkl -----F h k l = F h k 1. A preparation of our s-chitin crystals gives the transmission infrared spectrum shown in Figure 4. This spectrum which is of good resolution, closely resembles those reported for a-chitin from lobster tendon 4'7. In particular, the C : O region consists of three sharp absorption bands at 1656, 1621 and 1558cm -~. The occurrence of these bands allows us to confirm that these crystals are indeed from the ~ allomorph of chitin. In the present case, the crystals could not be orientated in order to be analysed by polarized infrared spectroscopy. Thus, the polarization of the three C - - O bands as well as those of the remaining spectrum could not be compared with those of lobster tendon chitin.
The results presented in this study are, we believe, the first attempts to prepare single crystals of s-chitin from dilute solution. The method which was selected is based on a general recipe which we are using for growing polysaccharide single crystals 1°. It involves a high temperature of crystallization and the use of low molecular weight polysaccharide fractions, selected in such a way that the length of the molecules corresponds roughly to the thickness of the crystal lamellae. Thus the difficulty of folding stiff polysaccharide chains at a lamellar surface is avoided and well developed polysaccharide crystals can be normally obtained within 1 h or so. To solubilize chitin, there are only a small number of solvents available and several of them are not stable at high temperature. This is the case for aqueous LiSCN which decomposes, leaving a solid sulphur-based precipitate. A compromise between the extent of solvent degradation and that of crystal growth had therefore to be made. This is why temperatures of only 200-220°C
lal
I
Figure 2 As in Figure 1. Inset." electron diffractogram properly orientated with respect to the specimen and corresponding to the circled area of the image
I
I
I
I
L
I
Figure 4 Fourier transform infrared spectrum of a thin film
of ~t-chitin crystalssimilar to those shown in Figures I and 2
a ~
200 •
O0
• .0
~
O0
0.. 080
b"
°-0 0"" B
•
0 0 O0
• 0
Figure 3 (A) Electron diffraction diagram obtained on one :t-chitin crystal with its length vertical. The pattern was obtained using the microdiffraction technique of Riecke and Ruska 15. (B) Schematic diagram corresponding to the pattern shown in (A)
Int. J. Biol. Macromol., 1992, Vol. 14, August
223
Single crystals o f e-chitin: J. E. Persson et al.
were used and the crystallization time had to be kept below 30 min. Such conditions give only the above thin and narrow e-chitin crystals which could not be improved substantially. One of the interesting results obtained in this study is that our chitin crystals displayed only diffraction spots which were consistent with the P212~2~ space groups. The diffraction diagrams of our crystals differed somewhat from those of the native e-chitin crystals encountered in the grasping spines of Sagitta where odd 0k0 reflections were observed. Several reasons may be given for such discrepancy. It may be that the crystal parameters of native Sagitta chitin are different from those of the above recrystallized samples. If this is the case and until a complete determination is achieved, a parallel chain structure for Sagitta chitin must still be considered. Another possibility is that the 'forbidden' reflections 100, 300, 500, 700 and 010 are very weak. They may not show in the present limited electron diffractograms where only the strongest diffraction spots are visible. Thus, unless the above crystals can be improved, and the exact origin of the 'forbidden' reflections in native s-chitin can be assigned, the crystallography of e-chitin will remain a subject of debate. Chitin is often compared with cellulose as these two structural polysaccharides have not only closely related molecular structures, but also m a n y morphological similarities. Single crystals of low molecular weight cellulose grown from solutions at relatively low temperatures, also display a needle-like morphology x6 which is very reminiscent of the one shown above for e-chitin. In both crystals, one of the equatorial diffraction spacings is perpendicular to the needle axis (110 for cellulose and 100 for e-chitin) and the corresponding crystallographic planes are therefore the growth planes of these crystals. Moreover, in both crystals, the polymer chain axis is perpendicular to the plane on which the crystals are lying. With cellulose, it was shown that a different morphology occurred when polymer fractions of higher molecular weight were recrystallized: in that
224
Int. J. Biol. Macromol., 1992, Vol. 14, August
case, fibrillar crystals resulted where the cellulose chain axis was coincident with the fibrillar axis 17. It remains to be seen whether chitin of high molecular weight, which is also known to adopt a fibrillar morphology 18 upon re-crystallization, will also follow this behaviour. If this is the case, the transition from a needle-like into a fibrillar morphology as a function of molecular weight would be an interesting topic to study. W o r k is presently under way in our laboratory along this line.
Acknowledgements One of us (J.E.P.) was a recipient of an 18 month fellowship from the Swedish National Board for Technical Development (S.T.U.).
References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Minke,R. and Blackwell, J. J. Mol. Biol. 1978, 120, 167 Meyer,K. H. and Pankow, G. W. Heir. Chim. Acta 1937,18,232 Clark, G. L. and Smith, A. F. J. Phys. Chem. 1936, 40, 863 Lotmar, W. and Picken, L. E. R. Experientia 1950, 6, 58 Carlstr6m,D. J. Biophys. Biochem. Cytol. 1957, 3, 669 Dweltz,N. E. Biochim. Biophys. Acta 1960, 44, 416 Pearson,F. G., Marchessault, R. H. and Liang, C. Y. J. Polym. Sci. 1960, 43, 101 Atkins, E. D. T., Dlugosz, J. and Foord, S. Int. J. Biol. Macromol. 1979, 1, 29 Rudall,K. M. in 'The Insect Integuments', (Ed. H. R. Hepburn ), Elsevier, Amsterdam, 1976, p. 21 Chanzy, H. and Vuong, R. in 'Polysaccharides, Topics in Structure and Morphology', (Ed. E. D. T. Atkins), Macmillan, London, 1985, p. 52 Cartier,N., Domard, A. and Chanzy, H. Int. J. Biol. Macromol. 1990, 12, 289 Domard, A. and Cartier, N. Int. J. Biol. Macromol. 1989,11,297 Hirano, S.,Ohe, Y.andOno, H. Carbohydr. Res. 1976,47,315
Le Poole, J. B. Philips Teehn. Runds 1947, 9, 33 Riecke, W. D. and Ruska, E. in 'Electron Microscopy', (Ed. R. Uyeda), Maruzen Co. Ltd, Tokyo, 1966, p. 208 Bulron,A. and Chanzy, H. J. Polym. Sci., Phys. Ed. 1978,16, 833 Quenin, I. and Chanzy, H. in 'The Structures of Cellulose', (Ed. R. H. Atalla), ACS Symposium Series, No. 340, American Chemical Society, Washington, 1987, p. 189 Yamaguchi,R., Arai, Y. and Itoh, T. Aoric. Biol. Chem. 1982, 46, 2379
(Received 31 October 1991; revised 20 February 1992)) Single crystals of ~t-chitin were grown by the addition of precipitants to dilute solutions of low molecular weight chitin fractions dissolved in aqueous LiSCN. At temperatures around 200°C, bundles of thin needle-shaped crystals were obtained. Each of these needles was an or-chitin single crystal, characterized by a spot electron diffraction pattern which could be indexed along the hkO reciprocal net corresponding to the Minke and Blackwell unit cell [a = 0.474 nm, b = 1.88 nm, c (fibre axis) = 1.032 nm, space group P212121]. In a crystal, the a* parameter was along the crystal axis and the b* perpendicular to it. Keywords: ~-Chitin; crystalstructure;i.r. spectroscopy
Introduction During the past 60 years, several studies have focused on the crystallography of ~-chitin 1-6. Among these reports, the most recent determination has been given by Minke and BlackwelP who have presented the latest molecular structure and atomic coordinates of this chitin allomorph. Their model which is refined in the space group P21212~, involves two antiparallel chitin chains per unit cell and one N-acetyl D-glucosamine (2-deoxy2-acetamido-fl, D-glucose) as the independent residue. One of the important features of the Minke/Blackwell model is that it contains a statistical distribution of two rotational conformations for the C H 2 O H moieties, each of them having an occupancy of one-half. This distribution has the advantage of allowing this group to be hydrogen bonded to the remainder of the structure. Such hydrogen bonding is required to confirm the results from infrared analysis, which indicate an absence of free O H groups in a-chitin 7. According to Minke and Blackwell 1, the perfect statistics of the two conformations at C H 2 O H is established only in large a-chitin crystals. In small ones, defects occurring especially at the crystal edges should, in particular, account for the 001 reflection which is forbidden in the P2~2~2: space group. In an electron diffraction study performed on the large chitin crystals 8 found in the grasping spines of Sagitta, Atkins et al. 9 were able to record a series of spot patterns corresponding to the 0kl and h01 sections of the reciprocal space of ~t-chitin. Quite remarkably, these diagrams displayed both 001 (with 1 odd) and 0k0 (with k odd) reflections, in contradiction to the expected diagrams where these reflections should be absent if the proposed P2~2a21 space group was correct. There are therefore conflicting indications concerning the crystal *Present address: Department of PolymerTechnology,Royal Institute of Technology,S-100 44 Stockholm, Sweden. tPresent address: Laboratoire d'Etudes des Mat6riaux Plastiques et des Biomat6riaux, URA CNRS No. 507, Universit6 Claude Bernard, 43, Bvd du 11 Novembre 1918, 69622 Villeurbanne cedex, France. :~To whom correspondenceshould be addressed. 0141-8130/92/040221-04 © 1992 Butterworth-HeinemannLimited
structure of a-chitin. Further clarification requires the preparation of a fair number of crystals of high perfection. From these, one should collect a series of meaningful data by infrared and ~3C C P / M A S n.m.r, spectroscopy, together with X-ray and electron diffraction. It is from the convergence of results from all these techniques, that the final description of the structure of a-chitin will emerge. As quantities of Sagitta grasping spines are not easy to collect, the preparation of single crystals of a-chitin was attempted. For this, a crystallization scheme was devised which follows the general rules devised in our laboratory for obtaining single crystals of polysaccharides ~°. The method involves the preparation of dilute solutions of low molecular weight chitin fractions, followed by their recrystallization at high temperature by the addition of a precipitant. The crystals obtained in such a way, together with their diffraction diagrams are described below.
Experimental Preparation o f chitin o f low degree o f polymerization ( D P )
Chitin is difficult to solubilize and its separation into low D P fractions is therefore hard to obtain. In this work, these fractions were prepared by N-acetylation of low D P chitosan (prepared as in the method described in previous papers 11,1z), using the selective re-acetylation method of Hirano et al. 13. For this, 70 mg of low D P chitosan (DPv = 3511) was dissolved in 1.4 ml of 10% (w/v) acetic acid. A volume of 7 ml of methanol was then added followed by 0.11 ml of acetic anhydride. The mixture was left at room temperature overnight during which a chitin gel formed. This gel was slurried in 100 ml of isopropanol for 24 h. The chitin precipitate was then isolated by centrifugation, re-dispersed in water and freeze-dried. This product, checked by infrared spectroscopy, was identical to an c~-chitin standard extracted from crab tendon. It was soluble in dimethyl acetamide saturated with LiC1 but not in water. In addition, its proton n.m.r, spectrum in D 2 0 / L i S C N gave the expected ratio of three protons at CH 3 to one at NH.
Int. J. Biol. Macromol., 1992, Vol. 14, August
221
Single crystals of o-chitin: J. E. Persson et al. Preparation of the e-chitin single crystals
Infrared spectroscopy
Several systems of solvent/precipitant were tested for the preparation of chitin crystals. Solutions of chitin were prepared in hexafluoroisopropanol, dimethyl acetamide saturated with LiC1, aqueous LiSCN and aqueous LiI. These solutions were recrystallized by adding precipitants such as water, alcohols, pyridine, acetone, dibenzyl ether, etc. at temperatures ranging from ambient to 250°C. The best crystals were prepared by adding aqueous poly(ethylene glycol) ( P E G 200) to a solution of low D P chitin in aqueous LiSCN, at temperatures between 200 and 220°C. Two samples of LiSCN were used: a batch from Fluka was used without purification; another batch from Aldrich did not give as good results as the Fluka LiSCN despite extensive dissolution, filtration and recrysallization. As LiSCN decomposes at high temperature to yield a sulphur-based precipitate which will contaminate the chitin crystals, a limited time is available for recrystallizing chitin. Our best crystallization recipe was as follows: 5 mg of low D P chitin was dissolved in 10 ml of an aqueous solution of LiSCN. This solution was prepared by adding water to LiSCN until its refractive index was 1.428 at 25°C. Then 0.5 ml of the chitin solution was poured into a thin-walled glass ampoule which was sealed and placed inside a small poly(tetrafluoroethylene) ( P T F E ) container, fitted with a P T F E screw cap. A small glass ball whose purpose was to break the ampoule was also put into the container. A measure of 2.5 ml of dilute aqueous P E G 200 (80% water v/v) was poured into the P T F E container which was screw-capped and inserted inside a steel bomb to which a small amount of dilute aqueous P E G was added to produce a counter pressure. The bomb was tightly sealed and immersed in an oil bath kept at 240°C. After 5 min at this temperature, the bomb was shaken vigorously to break the glass ampoule, which caused the chitin solution to be mixed with its precipitant, thus initiating the crystallization. The bomb was then immersed and put into a second oil bath heated to 220°C. After 30 min, the bomb was removed from the oil bath and cooled under tap water. It was opened and the suspension of chitin crystals was removed from the P T F E container. The crystals were washed by successive centrifugation, first in aqueous P E G 200 (80% water), then in water and finally in methanol. The crystals were then stored in methanol.
Infrared spectra were recorded in transmission mode with a resolution of 2 c m - 1. These spectra were recorded on a Perkin-Elmer 1720X FT-IR spectrometer using a microfocus accessory with a micro sample holder disc having a 500 #m aperture. For this, drops of chitin crystal suspensions were evaporated inside a poly(ethylene) cap to yield a thin film which could be handled at the surface of water and thus mounted across the hole of the sample holder.
Results and discussion A typical preparation of o-chitin crystals is shown in Figure 1. This electron micrograph displays a series of bundles each of them consisting of several dozens of needle-like crystals. Within a bundle, the crystals are tightly packed in the middle of the bundle but fan out in a typical sheaf morphology at both bundle tips. The bundles occur either isolated or associated with other bundles. In the latter case, the bundles radiate from a common centre. As seen in Figure 1, a typical bundle has a length of a few #m and a width of about 100 nm in its centre and around 500 nm at each of its ends. The crystals constituting the bundle have pointed tips. They run the full length of the bundle, but have widths that do not exceed 10-20 nm. Their thickness, deduced from images shadowed at an angle, is also in the order of 10 nm. When tested by selected area electron diffraction, a crystal bundle yields a poorly resolved electron diffraction pattern. This is illustrated in Figure 2 where the inserted diffractogram corresponds to the area circled in the image. Despite its poor resolution and substantial arcing, this diagram can be identified as corresponding to the hk0 diffraction arcs from o-chitin indexed along the unit cell defined by Minke and Blackwellt. The bundle axis is aligned along the a* axis of o-chitin and the b* axis is perpendicular to it. The arciffg of the pattern corresponds to the curvature taken by the crystals which fan out within the bundle. Patterns corresponding to one a-chitin crystal can be obtained by using the microdiffraction technique of Riecke and Ruska 15. This is exemplified in Figure 3 which also shows a schematic diagram and proper indexation. The pattern consists of spots which are mirrored around
Electron microscopy and electron diffraction analysis Drops of chitin crystal suspensions were deposited on carbon coated grids and allowed to dry. All electron microscopy and electron diffraction analysis was achieved with a Philips EM 400T electron microscope operated at 120 kV for diffraction purposes and 80 kV for imaging. Electron diffractograms were recorded on Mitsubishi electron image ( M E M ) films and images were produced with Kodak 4489 emulsion. The diffraction diagrams were obtained either with the conventional selected area technique of le Poole 14 or the microdiffraction method of Riecke and Ruska 15. The former method gives information on at least 1/xm 2 of the specimen whereas the latter is appropriate to obtain diffraction patterns on areas as small as a few tens of nm. Some of the electron diffractograms were calibrated by depositing the chitin crystals on carboncoated grids on which a thin layer of gold had been deposited by evaporation.
222
Int. J. Biol. Macromol., 1992, Vol. 14, August
Figure 1 Electron micrograph of a preparation of ~-chitin crystals prepared at 220°C by adding a precipitant (aqueous poly(ethylene glycol), PEG 200) to a dilute solution, of low molecular weight chitin in aqueous LiSCN
Single crystals of a-chitin." J. E. Persson et al. two orthogonal axes a* and b*, a* corresponding to the crystal axis and b* is perpendicular to it. The pattern in Fioure 3 has a resolution approaching 0.2 nm. It displays nine independent spots which can be seen in each quadrant. Quite interestingly, only even reflections (200 along a* and 020 and 080 along b* ) are seen on the two main axes of the diagram. Thus, this diffractogram is consistent with a pgg two-dimensional space group, which itself would meet the P212~2~ requirements provided that one could demonstrate that (a) c* is perpendicular to the a'b* plane, (b) Fhk I = Fffhkl -----F h k l = F h k 1. A preparation of our s-chitin crystals gives the transmission infrared spectrum shown in Figure 4. This spectrum which is of good resolution, closely resembles those reported for a-chitin from lobster tendon 4'7. In particular, the C : O region consists of three sharp absorption bands at 1656, 1621 and 1558cm -~. The occurrence of these bands allows us to confirm that these crystals are indeed from the ~ allomorph of chitin. In the present case, the crystals could not be orientated in order to be analysed by polarized infrared spectroscopy. Thus, the polarization of the three C - - O bands as well as those of the remaining spectrum could not be compared with those of lobster tendon chitin.
The results presented in this study are, we believe, the first attempts to prepare single crystals of s-chitin from dilute solution. The method which was selected is based on a general recipe which we are using for growing polysaccharide single crystals 1°. It involves a high temperature of crystallization and the use of low molecular weight polysaccharide fractions, selected in such a way that the length of the molecules corresponds roughly to the thickness of the crystal lamellae. Thus the difficulty of folding stiff polysaccharide chains at a lamellar surface is avoided and well developed polysaccharide crystals can be normally obtained within 1 h or so. To solubilize chitin, there are only a small number of solvents available and several of them are not stable at high temperature. This is the case for aqueous LiSCN which decomposes, leaving a solid sulphur-based precipitate. A compromise between the extent of solvent degradation and that of crystal growth had therefore to be made. This is why temperatures of only 200-220°C
lal
I
Figure 2 As in Figure 1. Inset." electron diffractogram properly orientated with respect to the specimen and corresponding to the circled area of the image
I
I
I
I
L
I
Figure 4 Fourier transform infrared spectrum of a thin film
of ~t-chitin crystalssimilar to those shown in Figures I and 2
a ~
200 •
O0
• .0
~
O0
0.. 080
b"
°-0 0"" B
•
0 0 O0
• 0
Figure 3 (A) Electron diffraction diagram obtained on one :t-chitin crystal with its length vertical. The pattern was obtained using the microdiffraction technique of Riecke and Ruska 15. (B) Schematic diagram corresponding to the pattern shown in (A)
Int. J. Biol. Macromol., 1992, Vol. 14, August
223
Single crystals o f e-chitin: J. E. Persson et al.
were used and the crystallization time had to be kept below 30 min. Such conditions give only the above thin and narrow e-chitin crystals which could not be improved substantially. One of the interesting results obtained in this study is that our chitin crystals displayed only diffraction spots which were consistent with the P212~2~ space groups. The diffraction diagrams of our crystals differed somewhat from those of the native e-chitin crystals encountered in the grasping spines of Sagitta where odd 0k0 reflections were observed. Several reasons may be given for such discrepancy. It may be that the crystal parameters of native Sagitta chitin are different from those of the above recrystallized samples. If this is the case and until a complete determination is achieved, a parallel chain structure for Sagitta chitin must still be considered. Another possibility is that the 'forbidden' reflections 100, 300, 500, 700 and 010 are very weak. They may not show in the present limited electron diffractograms where only the strongest diffraction spots are visible. Thus, unless the above crystals can be improved, and the exact origin of the 'forbidden' reflections in native s-chitin can be assigned, the crystallography of e-chitin will remain a subject of debate. Chitin is often compared with cellulose as these two structural polysaccharides have not only closely related molecular structures, but also m a n y morphological similarities. Single crystals of low molecular weight cellulose grown from solutions at relatively low temperatures, also display a needle-like morphology x6 which is very reminiscent of the one shown above for e-chitin. In both crystals, one of the equatorial diffraction spacings is perpendicular to the needle axis (110 for cellulose and 100 for e-chitin) and the corresponding crystallographic planes are therefore the growth planes of these crystals. Moreover, in both crystals, the polymer chain axis is perpendicular to the plane on which the crystals are lying. With cellulose, it was shown that a different morphology occurred when polymer fractions of higher molecular weight were recrystallized: in that
224
Int. J. Biol. Macromol., 1992, Vol. 14, August
case, fibrillar crystals resulted where the cellulose chain axis was coincident with the fibrillar axis 17. It remains to be seen whether chitin of high molecular weight, which is also known to adopt a fibrillar morphology 18 upon re-crystallization, will also follow this behaviour. If this is the case, the transition from a needle-like into a fibrillar morphology as a function of molecular weight would be an interesting topic to study. W o r k is presently under way in our laboratory along this line.
Acknowledgements One of us (J.E.P.) was a recipient of an 18 month fellowship from the Swedish National Board for Technical Development (S.T.U.).
References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Minke,R. and Blackwell, J. J. Mol. Biol. 1978, 120, 167 Meyer,K. H. and Pankow, G. W. Heir. Chim. Acta 1937,18,232 Clark, G. L. and Smith, A. F. J. Phys. Chem. 1936, 40, 863 Lotmar, W. and Picken, L. E. R. Experientia 1950, 6, 58 Carlstr6m,D. J. Biophys. Biochem. Cytol. 1957, 3, 669 Dweltz,N. E. Biochim. Biophys. Acta 1960, 44, 416 Pearson,F. G., Marchessault, R. H. and Liang, C. Y. J. Polym. Sci. 1960, 43, 101 Atkins, E. D. T., Dlugosz, J. and Foord, S. Int. J. Biol. Macromol. 1979, 1, 29 Rudall,K. M. in 'The Insect Integuments', (Ed. H. R. Hepburn ), Elsevier, Amsterdam, 1976, p. 21 Chanzy, H. and Vuong, R. in 'Polysaccharides, Topics in Structure and Morphology', (Ed. E. D. T. Atkins), Macmillan, London, 1985, p. 52 Cartier,N., Domard, A. and Chanzy, H. Int. J. Biol. Macromol. 1990, 12, 289 Domard, A. and Cartier, N. Int. J. Biol. Macromol. 1989,11,297 Hirano, S.,Ohe, Y.andOno, H. Carbohydr. Res. 1976,47,315
Le Poole, J. B. Philips Teehn. Runds 1947, 9, 33 Riecke, W. D. and Ruska, E. in 'Electron Microscopy', (Ed. R. Uyeda), Maruzen Co. Ltd, Tokyo, 1966, p. 208 Bulron,A. and Chanzy, H. J. Polym. Sci., Phys. Ed. 1978,16, 833 Quenin, I. and Chanzy, H. in 'The Structures of Cellulose', (Ed. R. H. Atalla), ACS Symposium Series, No. 340, American Chemical Society, Washington, 1987, p. 189 Yamaguchi,R., Arai, Y. and Itoh, T. Aoric. Biol. Chem. 1982, 46, 2379
