Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants (Nicotiana tabacum) to Comment

Harry. Hepworth, D. G. Et Al.

Description

Annals of Botany 81: 729–734, 1998Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants(Nicotiana tabacum) to Comment on the Molecular Structure of Cell WallsD.G.HEPWORTH*, J.F.V.VINCENT* and W.SCHUCH‹*Centre for Biomimetics, Earley Gate, TOB1, Reading UniŠersity, Reading RG6 2AT and ‹ Zeneca Seeds,Jealott’s Hill Research Station, Bracknell RG12 6EYReceived: 13 October 1977 Returned for revision: 5 January 1988 Accepted: 17 February 1998Lignin in the cell walls of woody tissue has a much lower crosslink density in tobacco wood than in tree wood. Thiscauses tobacco wood to show very different viscoelastic behaviour. With the aid of genetically modified plants, it isshown that the lignin in tobacco plant cell walls behaves in much the same way as a polymer solution. It exhibits bothrate stiffening and rate thinning behaviour due to the entangled nature of the lignin networks. The hydrophobicportions of lignin have a very low polymer chain crosslink density, hence entanglements make a significantcontribution. # 1998 Annals of Botany CompanyKey words: Tobacco plants, lignin, polymer solution, shear thinning, stress relaxation.Birkinshaw et al. (1989) identified two tan d peaks inINTRODUCTIONtemperature scans ranging from fi150 to 150 °C. However,Plant cell walls are viscoelastic and the Young’s moduli are in all cases tan d never fell below 0–01, and above fi50 °Crate dependent (Preston, 1974; Wainwright et al., 1976; was always above 0–02. Thus ...

Informations

• Publié par: Zyon
• Nombre de lectures: 32
• Langue: English

Extrait

Annals of Botany81: 729–734, 1998

Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants (Nicotiana tabacum) to Comment on the Molecular Structure of Cell Walls

D. G. HE P W O R T H*,J. F. V. VI N C E N T*and W. SC H U C H† * Centrefor Biomimetics, Earley Gate, TOB1, Reading Uniersity, Reading RG6 2AT and†Zeneca Seeds, Jealott’s Hill Research Station, Bracknell RG12 6EY

Received :13 October 1977Returned for revision: 5 January 1988Accepted :17 February 1998

Lignin in the cell walls of woody tissue has a much lower crosslink density in tobacco wood than in tree wood. This causes tobacco wood to show very different viscoelastic behaviour. With the aid of genetically modified plants, it is shown that the lignin in tobacco plant cell walls behaves in much the same way as a polymer solution. It exhibits both rate stiffening and rate thinning behaviour due to the entangled nature of the lignin networks. The hydrophobic portions of lignin have a very low polymer chain crosslink density, hence entanglements make a significant contribution.#1998 Annals of Botany Company

Key words:Tobacco plants, lignin, polymer solution, shear thinning, stress relaxation.

Birkinshawet al. (1989) identified two tanδpeaks in I N T R O D U C T I O N temperature scans ranging from®150 to 150°C. However, Plant cell walls are viscoelastic and the Young’s moduli are in all cases tanδnever fell below 0±01, and above®50°C rate dependent (Preston, 1974; Wainwrightet al always above 0 ; was., 1976 ±02. Thus there could easily be a large Dinwoodie, 1989). Stress relaxation experiments on whole number of overlapping transitions which are not separable. Nitella opacacells (Haughton, Sellen and Preston, 1968) This is not surprising given the complex structure of most showed that there is a 13% decay in the initial value of biological materials. Birkinshawet al. (1986) reported that stress after 300 sec. Beyond this time, relaxation is negligible. the responses of dried woods are not species specific. Kelley This decay is temperature insensitive, which is characteristic et al. (1987) found that moisture greatly affects the of crystalline polymers. Therefore the relaxation properties temperature at which the transitions occur. Increasing the ofNitellacell walls are controlled by cellulose, hemicellulose water content reduced the storage modulus (E«) at all and pectin. Wood from trees has similar relaxation temperatures and increased tanδ. From their data, it also properties, but the relaxation is temperature sensitive. seems that some of the tanδpeaks are no longer visible at Above 20°C there is a marked increase in rates of creep and a moisture content of 30%. DMTA is a useful tool for stress relaxation (Dinwoodie, 1989). This can be attributed investigating properties of secondarily-thickened cell walls, to the presence of lignin constituents which are above their and along with other methods, is beginning to show just glass transition temperatures. The water content is known how complex the molecular architecture is. to affect the relaxation and creep properties of wood In this investigation, secondary xylem tissue from tobacco (Kollman, 1962). Lignin is a complex polymer composed of plants (Nicotiana tabacum‘ Samsun ’)was used, because a a number of different monomers. Not only does the number of genetic modifications have been carried out to composition of lignin vary from plant to plant, it also varies produce plants with altered lignin, thus providing a useful across individual cell walls (Takabeet al experimental tool for investigating the properties of the ; Terashima,., 1986 Fukushima and Takabe, 1986; Hibinoet al., 1994). This lignin matrix within the cell wall. Previous work on lignified means that relaxation properties of woody tissues from xylem tissue (wood) from tobacco plants indicated that the various plants could be quite different. content of lignin in tobacco plants is similar to that in tree A number of different methods have been used to wood (40–45%), but it swells with water to a much greater investigate viscoelastic properties. One such method is extent (Hepworth, 1997). This suggests that tobacco plant dynamic mechanical thermal analysis (DMTA) which has lignin has a much lower crosslink density than that of trees. been used to investigate wood from different tree species DMTA reinforces this hypothesis because the value of tan (Birkinshaw, Buggy and Henn, 1986; Kelley, Rials and δat 50°C in tobacco wood is much higher than in wood Glasser, 1987; Birkinshaw, Buggy and Carew, 1989). The from most trees (in both dry and wet states) indicating that specimens were tested in shearing and bending modes in the more of the molecules are undergoing a glass transition air dried state. Most species with wood show a number of (changing from glassy to rubbery). Thus the viscoelastic overlapping tanδ behaviour of tobacco wood should differ to that of tree peaks which are thought to represent the glass transitions of hemicellulose and lignin components wood. within the cell wall matrix. Kelleyet al. (1987) and Two types of genetically modified tobacco plants were 0305-7364}98}06072906 $25.00}0 bo980619#1998 Annals of Botany Company

730Hepworthet al.—Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants compared with control plants. The first contained antisenseδwere measured, and the data stored on disk. Specimens cinnamyl alcohol dehydrogenase (CAD) DNA. This is onewere tested as a dual cantilever beam (clamped at both ends of the last enzymes in the pathway leading to ligninand in the centre) at a fixed frequency of 1 Hz. They were synthesis. The antisense DNA blocks the conversion of then subjected to a temperature scan which increased at a − phenolic aldehydes to phenolic alcohols (Halpinet al., 1992 ;rate of 5°C min " . Specimens were tested either fully Knight, Halpin and Schuch, 1992) and is known to result inhydrated or dry. Wet specimens were kept hydrated during the production of lignin with a lower crosslink densitythe test by wrapping wet paper towelling round the inside of (Halpinet al., 1994 ; Hepworth, 1997). The second containedthe head which could be made air tight. For these specimens antisense cinnamyl co-enzyme reductase (CCR) DNA. Thisthe temperature scan was limited to 20–95°C ;beyond this is the first enzyme in the phenolic pathway which is specific they lost water too rapidly. Specimens were used only once. to lignin formation. Blocking its production seems to reduceThe clamping was arranged so that specimens were tested at the production of lignin (O’Connell and Schuch, pers.a low span to depth ratio, producing a large amount of comm.). sheardeflection ; this should improve resolution of the lignin The aim of this paper is to show how a combination ofmatrix properties. different mechanical tests can be used to determine theTo investigate if there are components of the lignin viscoelastic properties of secondarily-thickened xylem cellcomplex which are not covalently crosslinked to other cell walls of tobacco plants. With the aid of data from genetically wall components, but are just held in place by hydrophobic modified varieties, a theory regarding the microstructure ofinteractions, wood was coarsely chopped (1mm cubes), the cell wall matrix will be proposed.washed with cold methanol for 1 h and then extracted with boiling methanol for 1h. The weight of dried extract removed from the wood was recorded and the constituents were identified by re-dissolving the extract in alcohol and M A T E R I A L SA N DM E T H O D S running it through an HPLC. The column contained C 18 A number of mechanical tests were performed on secondaryphenyl, and the mobile phase used was a mixture of equal xylem (woody tissue) which had been separated from the parts of acetic acid, methanol and water. The flow rate was − pith and cortex. The wood was cut into 5 cm¬1 ml min5 mm strips " . The absorbence spectra of the extracted com- along the length of the grain, and the strips were equilibrated pounds were recorded. The values and shapes of the to the required relative humidity. Wood which was required absorbence curves and the residency times could be fully dry was heated at 120°C for 8h. Tensile tests were compared with those of various known compounds, such as performed on wood from control and antisense CAD and coniferyl alcohol, coniferyl aldehyde, sinapyl alcohol, and CCR plants at different rates and temperatures using an sinapyl aldehyde. Instron testing machine. Specimens were wrapped in R E S U L T S clingfilm after clamping to prevent moisture exchange. For fully hydrated specimens, the clingfilm wrapping was filled Figure 1 shows that, for fully hydrated control wood, with the treatment solution (distilled water or 10% Triton), increasing the rate of deflection increased the load at and for dry specimens it was filled with silica gel and the 0±15 mmdisplacement until a peak was reached at a specimen was kept hot. The specimens were tested to a − deflection rate of 5mm min " . After this there was a very maximum strain of only 0±Each specimen was tested at5 %. − pronounced dip in the curve between 5mm min " and − eight different loading rates, ranging from 0±01 mm min " to − 20 mm min " . Here the force actually decreased with − 100 mm min " min, with a recovery period of at least 30 between each test. Some specimens were also tested at different temperatures. All specimens were cut to the same 2 dimensions (5cm long¬2±5 mmwide¬thick). For1 mm Control each loading event, the load at 0±15 mm displacement was Antisense recorded. The temperature could be changed by inserting a small heating pad inside the clingfilm wrapping. A thermal probe inside the wrapping recorded the temperature. Some specimens were pre-treated by soaking in 10% Triton for 1 48 h and testing in a fully hydrated state, or by drying before testing. Stress relaxation experiments were performed by stretching the specimens to a specified strain at a rapid rate − (200 mm min " ) and then allowing the stress to relax over time. The apparatus was the same as that described above. The same specimen could be tested at different temperatures 0 –3 –2 –10 1 2 with a long period of recovery (5 h) between each test. –1 log speed(mm min) Dynamic mechanical testing analysis (DMTA, Polymer 10 Labs) involved vibrating the specimen at a particular F. 1. Force at 0±15 mmdisplacement during tensile tests of fully frequency and temperature both of which were computer hydrated tobacco wood performed at different rates of deformation. controlled. Storage modulus (E«), loss modulus (E§) and tan Nine samples from three plants were used.

Hepworthet al.—Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants731 2 1.2 Control Untreatedcontrol CCR modified Triton treated control 1.0

1

0.8

0.6

0 0.4 –3 –2 –10 1 2 –3 –2 –10 1 2 –1 –1 ed (mm min) log 10 spe log )speed (mm min 10 F. 2. Force at 0±15 mm displacement during tensile tests of cinnamyl F. 4. Force at 0±15 mm deflection during tensile tests of control wood co-enzyme reductase (CCR) modified wood performed at different pre-treated with 10% Triton performed at different rates of de- rates of deformation. Eight samples from two plants were used. formation. Nine samples from three plants were used.

6 Control 5 Antisense

4

3

2

1

0 –3 –2 –10 1 2 3 –1 log speed(mm min) 10

F. 3. Force at 0±15 mmdisplacement during tensile tests of dry control and antisense wood performed at different rates of deformation. Nine samples from three plants were used.

increasing rate of deflection. This dip in the force}rate curve occurred for every specimen tested and is not an artefact caused by averaging data. The significance of the dip can be shown by recording the difference in the value of force between the peak and the trough at higher rates of deformation. Analysis of variance was performed to test whether these differences were significantly different from zero. For both the wet control and wet antisense wood the differences were significant (P¯0±0005 and 0±01, respect- ively). In antisense wood this peak and trough in the curve was flattened out. At slow and fast rates of deflection the force values were the same for control and antisense wood, − but at intermediate rates (5 mm min " ) the antisense wood required a lower force. Exactly the same curve was produced if the same piece of wood was allowed to recover for several hours and then retested under the same conditions. In CCR wood the peak and trough in the curve had virtually disappeared and the difference was not significant (Fig. 2). The values of force at all rates of deflection were reduced.

1.1

1.0

0.9

0.8

0.7 0 1020 30 40 50 Temperature (°C) F. 5. The effect of temperature on the force at 0±15 mm displacement during tensile tests in which the specimen was deformed at a rate of − 1 mm min " . The same specimen was tested at different temperatures to produce each curve.

Drying control and antisense wood increased the required load at 0±15 mm deflection and made the peak and trough in the curve more pronounced and more significant (control − and antisense,P¯1±98¬10 ' ; Fig. 3). There was no longer a detectable difference between control and antisense wood. Pre-treatment with 10% Triton reduced the required force (at 0±15 mm displacement) at slow rates of deflection, but had little effect at high rates of deflection (Fig. 4). The peak and trough became more significant in control wood − (P¯7±56¬10 ' ) and not significant in antisense wood (P¯ 0±14). Cooling control wood had a very similar effect on the force}rate curve to pre-treatment with Triton. Changing the temperature also produced a peak and trough in the force}temperature curve; this is shown in Fig. 5 where the force at 0±15 mmdisplacement is shown as a − function of temperature (deformation rate of 1 mm min " ). Only control wood was investigated in this way. Stress relaxation experiments showed that for wet control

732Hepworthet al.—Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants 1.0 0.3 A

0.8

0.2

0.6 2°C 0.1 24°C Antisense 0.4 35°C 53° C Control 0.2 0.0 0 1 2 3 20 40 60 80100 120 log time(s) Temperature (°C) 10 9.2 F. 6. Stress relaxation of wet control wood after 0±deformation5 % B − at 200 mm min " , as a function of time. The same specimen was tested at different temperatures. 9.0 Control 8.8 1.2

1.0

0.8

0.6

0.4

0.2 0 1 2 3 4 log time(s) 10 F. 7. Stress relaxation of dry control wood after 0±deformation5 % − at 200 mm min " , as a function of time. The specimen was tested at 53°C.

wood the rate of relaxation is temperature sensitive (Fig. 6). When both force and time are plotted on a logarithmic scale, the relaxation curve shows two stage relaxation. Within the first 10 sec after loading stops, the force relaxes rapidly. After 10sec it levels off, and for the next 30sec there is very little relaxation. The rate of relaxation then begins to increase again, provided that the temperature is greater than 24°C. The second period of relaxation is very sensitive to temperature changes. At 2°C it is hardly noticeable, but at 53°C all the remaining stress relaxes within 6 min. In dry control wood all the stress can still be relaxed within 30 min (Fig. 7). In wet wood in a high shear dual cantilever test, tanδis higher at lower temperatures (30°C) in antisense wood (Fig. 8 A).At 90°C there was no significant difference between tanδof control wood and tanδof antisense wood. The storage modulus is always lower in antisense wood, but the

8.6

8.4

Antisense

8.2 20 40 60 80 100 120 Temperature (°C) F. 8. The results of DMTA carried out on wet control and antisense wood at a frequency of 1 Hz. The mean curves of ten samples of each type of wood are shown along with lines indicating 95% confidence limits of the data. A, Tanδ; B, E«.

difference between control and antisense wood decreases at higher temperatures (Fig. 8B). Drying increases the storage modulus and reduces tanδ slightly over the whole temperature range (Fig. 9B). The loss moduli also increase slightly during drying in control and antisense wood. The biggest reduction in tanδoccurs at lower temperatures. After drying the wood, a number of obvious tanδpeaks can be seen at 50 and 100°C (Fig. 9 A).There is no longer any difference in the storage moduli between control and antisense wood. However, tanδand the loss modulus are lower in antisense wood between 30 and 100°A).C (Fig. 9 The filtrate produced during the grinding of cell wall material was found to contain chlorophyll and a number of unidentified compounds. Nothing relating to lignin was found. Boiling cell wall material with methanol resulted in the extraction of coniferyl alcohol, coniferyl aldehyde, sinapyl alcohol and sinapyl aldehyde from antisense wood, and coniferyl alcohol from control wood. A number of unidentified high molecular weight compounds were also extracted. 2±of the initial dry weight of antisense wood2 % was removed in the extract and 1±of the initial dry2 % weight of control wood.

Hepworthet al.—Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants733 0.12 hump is reduced in significance in antisense wood suggests A that there are fewer entanglements. This could be because the entangling chains are shorter in antisense wood due to the reduced crosslink density. This is further supported by 0.10 the lack of a hump in the curves for CCR wood, which contains less lignin. The main effects (the hump in the force}rate curve and the 0.08 step in the stress relaxation curve) cannot be explained by water and hydrophobic associations because both the force}rate curves and the stress relaxation curves still Antisense 0.06 exhibit the same characteristics when they are fully dry; however, hydrophobicity does have a role to play. Treatment Control with Triton and cooling both reduce the force at lower 0.04 temperatures and reduce the significance of the hump. This 0 100200 300 can be explained by the fact that both Triton treatment and Temperature (°C) cooling reduce hydrophobicity. If the hydrophobic regions 9.5 B are also the regions with entangled polymer solution Antisense properties, then allowing water into these regions will 9.4 Control reduce entanglement friction and flatten the hump. The fact that the force}rate experiments are completely 9.3 reproducible for any particular specimen of wood rules out the presence of damage accumulation processes. 9.2 The polymer entanglement properties may be produced by lignin while the cellulose, hemicellulose and pectin 9.1 provide a fully crosslinked backbone structure with more normal elastic and plastic properties. How does the theory 9.0 of entanglement fit with what we know about the micro- structure of plant cell walls in general? Erinset al. (1976) 8.9 deduced a model for the structure of the matrix of wood cell 0 100200 300 walls based on various chemical extraction techniques. They Temperature (°C) concluded that lignin has a globular structure, with F. 9. The results of dynamic mechanical thermal analysis carried out hemicellulose chains linking several globules together. Each on dry control and antisense wood at a frequency of 1 Hz. Mean curves globule consists of about 20 tightly crosslinked phenyl- of ten samples of each type of wood are shown along with lines propane units and has a molecular weight of 4000. They indicating 95% confidence limits of the data. A, Tanδ; B, E«. suggested that many of these globules could be found in association, forming larger aggregations with molecular D I S C U S S I O N weights of up to 100000. In tobacco wood cell walls the The stress relaxation curve at 53°C is similar to that of a basic globules may not be so crosslinked internally, or there highly entangled polymer solution, with the step or flat may be a more branched structure with ‘dangling ’side region corresponding to the entanglement periodicity chains. These types of molecular structures could provide (Heinrich, Straube and Helmis, 1988; Graessley, 1993; the basis for entanglements. Grosberg and Khokhlov, 1994). It is very different to that of Reanalysis of the data of Kelleyet al. (1987), who wood from trees, where there is no step or plateau in the measured the dynamic mechanical moduli of various types curve and only about 30% of the stress is ever relaxed of wood at different temperatures, shows that there is a (Wainwrightet al hump in the stress ., 1976). This indicates that in tobacco }temperature curve for spruce wood. The wood certain components of the cell wall have a lower contribution of this peak to the total modulus is much crosslink density than in trees, allowing large amounts of smaller than for tobacco wood. However, this is to be permanent deformation when load is applied over a long expected, since the cell walls in tree woods have a much time scale. If a particular component of the cell wall is higher density than the matrix in tobacco wood cells. behaving as an entangled polymer solution, then this could DMTA analysis was carried out at temperatures greater also explain the hump in the force}rate and force} than 25°C ;thus the entanglement peak would have just temperature curves. The hump may represent competition been passed and normal glass transitions of the crosslinked between two processes. The ascending limb could represent backbone structures were observed. In DMTA tests of wet increasing entanglement friction with neighbouring polymer wood, tanδforms an almost flat straight line with increasing chains, and the descending limb could represent shear temperature (Fig. 8A). This means that E«and E§are thinning. Such a mechanism would also be sensitive to decreasing by almost the same percentage with increasing temperature, and this would explain the hump in the temperature. This is a characteristic normally associated force}temperature curve. This rate–temperature equivalence with polymer solutions, and supports the theory that the means that this effect is truly a thermodynamic one and is hydrophobic regions of the wood behave in similar ways to not the result of microstructural effects. The fact that the entangled polymer solutions. For true melts, however, E§is

734Hepworthet al.—Using Viscoelastic Properties of the Woody Tissue from Tobacco Plants higher than E«; this is not the case in tobacco wood (Fig. or reduce the lignin content significantly affect those results 8 B).Therefore there must be a covalently linked backbone which have been ascribed to entanglement. with either regions of totally uncrosslinked chains or entangling side chains hanging off the backbone into the spaces. Drying wood increases E«and E§but reduces tanδ L I T E R A T U R EC I T E D at 30°C (Fig. 9A and B). The contribution of an Birkinshaw C, Buggy M, Henn GG. 1986.Dynamic mechanical analysis entanglement effect to the overall modulus has been reduced. of wood.Journal of Material Science Letters5: 898. This can be explained by closure of the hydrophilic regions Birkinshaw C, Buggy M, Carew A. 1989.Thermo-mechanical behaviour of wood and wood products.Journal of Material Science24: and an increase in hydrogen bonding within them. This will 359–362. increase the modulus of the wood, and thus reduce the Dinwoodie JM. 1989.Wood natures cellular, polymeric fibre-composite. percentage contribution that the unchanged hydrophobic Vermont :Institute of Metals. regions make with their melt-like characteristics. This can Erins P, Cinite V, Jakobson M, Gravitis J. 1976.Wood as a be seen by looking at the force}deflection rate curves for wet multicomponent crosslinked polymer system.Journal of Applied Polymer Science:Applied Polymer Symposium28: 1117–1138. and dry wood (Fig. 3). The difference between the force at Graessley WW. 1993.Viscoelasticity and flow in polymer solutions and − − 5 mm min " and 100mm min " is a smaller percentage of concentrated solutions. In:Physical properties of polymers, 2nd − the value of force at 100 mm min " in dry wood. However, edn. Washington, DC: ACS Professional Reference Book, the magnitude of this difference increases with drying. This American Chemical Soc. is probably because drying creates some new entanglement Grosberg A Yu, Khokhlov AR. 1994.Statistical physics of macro- molecules: American Institute of Physics.. Woodbury, New York regions where hydrogen bonding is low. This is supported Halpin C, Knight ME, Grima-Pettenati J, Goffner D, Boudet A, Schuch by the fact that in the wet state the force}deflection rate W. 1992.Purification and characterisation of cinnamyl alcohol curve for antisense wood does not show much of an dehydrogenase from tobacco stems.Plant Physiology98: 12–16. entanglement peak, whereas it does after drying (Figs 1 and Halpin C, Knight ME, Foxon GA, Campbell MM, Boudet AM, Boon JJ, 3) when the less crosslinked or shorter chains have been Chabbert B, Tollier M-T, Schuch W. 1994.Manipulation of lignin quality by down regulation of cinnamyl alcohol dehydrogenase. brought into a close enough association for entanglements Plant Journal6: 339–350. to occur. Haughton PM, Sellen DB, Preston RD. 1968.Dynamic properties of The idea that certain regions within the cell walls have a Nitellacell walls.Journal of Experimental Botany19: 1–12. low crosslinking density is also supported by the fact that Heinrich G, Straube E, Helmis G. 1988.Rubber elasticity of polymer significant quantities of lignin components can be extracted networks ::Theories. InAdances in polymer science, 85, polymer from the cell walls of tobacco wood using hot alcohol, physics.London :Springer-Verlag. Hepworth DG. 1997.The mechanical properties of tobacco plants with which could not break covalent bonds. The quantities modified lignin.PhD Thesis. The University of Reading, UK. extracted are higher than those extracted from tree wood by Hibino T, Shibata D, Ito T, Higuchi T, Pollet B, Lapierre C. 1994. the Brauns method (usually 0±1–1±5 %),even though this Chemical properties of lignin fromAralia cordata.Phytochemistry method does not control for compounds which may have 37: 445–448. been located in the remains of the living protoplast and in Kelley SS, Rials TG, Glasser WG. 1987.Relaxation behaviour of the amorphous component of wood.Journal of Material Science22: living ray cells (Sarakanen and Ludwig, 1971). Also, our 617–624. extraction was much shorter than the Brauns procedure; Knight ME, Halpin C, Schuch W. 1992.Identification and charac- therefore we would expect to obtain larger quantities of terisation of cDNA clones encoding cinnamyl alcohol dehydro- extract if the Brauns method was followed. genase from tobacco.Plant Molecular Biology19: 793–801. There are a number of other possible explanations for Kollman F. 1962.Rheology of wood.Materialpru X fung4: 313–319. Preston RD. 1974.The physical biology of plant cell walls. London: these results, none of them are particularly convincing. For Chapman and Hall. example, adiabatic changes could produce a rate thinning Sarkanen KV, Ludwig CH. 1971.occurrence, formation,Lignins ; effect, but in this case the effects of drying should be much structure and reactions.New York: Wiley-Interscience. greater while substances such as Triton should have a Takabe K, Fujita M, Harada H, Saiki H. 1986.Lignification process in negligible effect. CryptomeriaElectron microscope observation of lignintracheid : skeleton of differentiating xylem.Research Bulletins of the The evidence described here indicates that certain portions Experimental Forests43: 783–788. of the cell wall matrix have a very low polymer chain Terashima N, Fukushima K, Takabe K. 1986.Heterogeneity in formation crosslink density which allows entanglement properties to of lignin.Holzforschung40: 101–105. become significant. This region probably corresponds to Wainwright SA, Biggs WD, Currey JD, Gosline JM. 1976.Mechanical lignin ‘globules ’,since genetic manipulations which change design in organisms. London: Edward Arnold.