Characterization of Cross-Links between Cellulose Microfibrils, and

Plant CellPhysiol. 41(4): 486-494 (2000)JSPP © 2000

Characterization of Cross-Links between Cellulose Microfibrils, and TheirOccurrence during Elongation Growth in Pea Epicotyl

Takeshi Fujino1, Yoshiaki Sone2, Yasushi Mitsuishi3 and Takao Itoh1

1 Wood Research Institute, Kyoto University, Uji, Kyoto, 611-0011 Japan2 Department of Food and Nutrition, Faculty of Science of Living, Osaka City University, Osaka, 558-8585 Japan3 National Institute of Bioscience and Human Technology, AIST, 1-1 Higashi, Tsukuba, Ibaraki, 305-0046 Japan

The occurrence and chemical nature of the cross-linksbetween cellulose microfibrils in outer epidermal cell wallsin Pisum sativum cv. Alaska was investigated by rapid-freezing and deep-etching techniques coupled with chemicaland enzymatic treatments. The cell wall in the elongatingregion of epidermal cells was characterized by the absenceof the cross-links, while in the elongated region, the cellwall was characterized by the presence of cross-links. Thecross-links remained in the cell wall of the elongated regionafter treatment with SDS electrophoresis sample buffer andtreatment with 4% potassium hydroxide. After treatmentwith endo-l,4-/?-glucanase, which fragments xyloglucan,the cross-links were remarkably reduced from the cell wallof the elongated region. The endoglucanase treatment alsoreduced immunogold labeling of xyloglucan in the cellwall. The endoglucanase hydrolysate from the cell wallfraction of the elongated region gave spots of oligosac-charides in thin layer chromatography, which were identi-cal to the spots of xyloglucan oligosaccharides produced byxyloglucanase from both the cell wall fraction and tama-rind xyloglucan. These results indicate that the cross-linksare made of xyloglucan. We discussed the possibility ofcross-links involved in the control of mechanical propertiesof the cell wall.

Key words: Cell elongation — Cross-links — Deep-etching— Epidermal cell — Pea (Pisum sativum cv. Alaska) —Xyloglucan.

Recent advances in the study of plant cell wall haveemphasized the importance of the cell wall in the growthand differentiation of plant cells (Hahn et al. 1989, Roberts1990). The structure of primary cell wall in dicotyledonousplants have been proposed to be consisted of a network ofCMFs cross-linked by xyloglucan, which is embedded in agel matrix of pectic polysaccharides (Carpita and Gibeaut1993). CMF has a crystalline structure consisting of linear1,4-/?-glucan chains and has a high tensile strength. Xylo-glucan contains a 1,4-/?-glucan backbone with xylose side

Abbreviations: CMF, cellulose microfibril; RFDE, rapid-freezing and deep-etching techniques; TLC, thin layer chromatog-raphy.

chains and additional galactosyl and fucosyl-galactosyl-residues along the backbone (Hayashi 1989). In order todissociate xyloglucan from the CMFs, strong alkali such as4 M solution of potassium hydroxide is required yieldingto disrupt the crystalline structure in CMF. This suggeststhat xyloglucan not only attach to the surface of CMF butalso extend into crystalline region of CMF. Whitney et al.(1995) found that cross-links were formed between cellu-lose ribbons produced by Acetobactor in the presence ofxyloglucan in culture medium. These findings indicate thatxyloglucan tightly binds to and cross-links to CMFs form-ing the principal load-bearing framework (Fry 1986, Tal-bott and Ray 1992). It is considered that the structuralchanges in the cellulose-xyloglucan framework is requiredduring cell elongation. This process may occur by theaction of enzymes or proteins related to cell elongation(Cosgrove 1997, Fry 1995, Hoson 1993, Hayashi 1991,Nishitani 1997). Although there is suggestion that xyloglu-can cross-links control the loosening of the cell wall (Taiz1984), the changes in the cellulose-xyloglucan frameworkduring elongation growth are poorly understood and havenot been well visualized.

RFDE is the only method to visualize cross-links be-tween CMFs in the cell wall. The cross-links are supposedto be composed of xyloglucan, based on an application ofsequential chemical extraction coupled with RFDE (Mc-Cann et al. 1990, Itoh and Ogawa 1993). Although chemi-cal treatment is effective in extracting cell wall components,it is relatively non-selective and causes modification of cellwall components such as the swelling of CMFs induced bystrong alkali. Enzymatic treatment is more selective thanchemical treatment when it comes to removing certain cellwall components. Endoglucanases derived from plant andfungi fragment xyloglucan and produce xyloglucan oligo-saccharides (Hayashi and Maclachlan 1984). These areavailable in high purity and are relatively specific. The useof endoglucanase instead of chemicals may provide addi-tional evidences whether the cross-links are composed ofxyloglucan in the cell wall.

We report the occurrence of cross-links in the cell wallduring and after cell elongation in the epidermal cell wallsof pea epicotyl. We focus on the characterization of thecross-links. That is, the chemical nature of cross-links wasexamined by comparing RFDE images before and aftervarious kinds of chemical and enzymatic extraction of cell

486

Dow

nloaded from https://academ

ic.oup.com/pcp/article/41/4/486/1816032 by guest on 04 February 2022

Xyloglucan cross-links in the cell wall 487

wall components. We discuss the possible role of the cross-links in the cell wall in relation to the control of mechanicalproperties during cell elongation.

Materials and Methods

Plant materials—Segments of the elongating and elongatedregions in third internode were prepared from 8-day old etiolatedpea seedling (Pisum sativum cv. Alaska) as described by Fujinoand Itoh (1998). Briefly, the elongating region is a 0-3 mm belowthe hook base and the elongated region is a 0-3 mm in the upperside of the node.

Immunogold labeling electron microscopy—The segmentswere obtained from elongating and elongated regions. The seg-ments from elongated region were subjected to EDTA extraction(0.1 M EDTA [pH7.0] at 85°C for 6h, the extraction solutionwas changed every 2 h) and followed by endoglucanase extraction(50 milliunits ml"1 cellulase [endo-l,4-/?-glucanase] from Tricho-derma sp. purchased from Megazyme, Australia in 50 mM sodiumacetate buffer [pH5.0] at 40°C for 24 h). Before the endo-glucanase treatment, the segments were cut half longitudinallywith surgical blade. The fixation, dehydration, embedding andsectioning of tissues for immunogold labeling were done accord-ing to Fujino and Itoh (1998).

The sections were incubated with blocking solution of oval-bumin (OA)-PBST (1% [w/v] ovalbumin in 10 mM phosphatebuffer [pH7.2] containing 500 mM sodium chloride, 0.1% [w/v]sodium azide and 0.5% [v/v] Tween 20) for 30min at roomtemperature. Sections were transferred into 20,ul of diluted pri-mary antibody (1 : 10 to 1 : 10,000 dilutions in OA-PBST of anti-xyloglucan [anti-xyloglucan heptasaccharides] antibody and a 1 :50 dilution of anti-xylan [anti-l,4-/?-D-xylose] antibody, Genosis,U.K.) and incubated for 2 h at room temperature. After the sec-tions had been washed with a continuous stream of PBST, 20 /ul ofsecondary antibody (anti-rabbit IgG conjugated 10 nm gold par-ticle diluted 1 : 20 in OA-PBST) was applied for 2 h. Sections werethen rinsed with PBST for 30 s followed by distilled water andwere stained with saturated uranyl acetate in water.

RFDE electron microscopy—The segments obtained fromthe elongating and elongated regions were extracted with 0.1 MEDTA (pH7.0) at 85°C for 6h (the extraction solution waschanged every 2 h). These segments were thoroughly washed withdistilled water. The segments from elongated regions were sub-jected to further extraction; boiling with SDS electrophoresissample buffer containing 2% (w/v) SDS and 10% (v/v) /?-mer-captoethanol (Laemmli 1970) for 30 min; treatment with 4%(w/v) potassium hydroxide containing 0.1% (w/v) sodium boro-hydride at 20° C for 8 h in a sonic bath (the extraction solution waschanged every 2h); and incubation with 50 milliunits mP 1 en-doglucanase in 50 mM sodium acetate buffer (pH 5.0) for 24 h at40°C. Before the endoglucanase treatment, the segments were cuthalf longitudinally with surgical blade. The segments, thoroughlywashed with distilled water, were rapid-frozen with liquid heliumand deep-etched, then replication was carried out as described inFujino and Itoh (1994). All pictures of deep-etched replicas werepresented as negative image.

Fractionation of cell wall, enzymatic hydrolysis and analysisof hydrolysates—The segments obtained from elongated regionswere frozen in liquid nitrogen and homogenized with a mortarand pestle in 80% (v/v) ethanol. The homogenate was extractedtwice with chloroform : methanol ( 1 : 1 , v/v) and twice with ace-tone and washed with distilled water. This material was treated

with 0.1 M EDTA (pH7.0) at 85°C for a total of 6 h (the ex-traction solution was changed every 2 h), then washed 5 times withdistilled water. The residue was then treated with 4 units ml"1

porcine pancreas a-amylase (type I, Sigma) in 25 mM sodium ac-etate buffer pH 6.5 for 2 h at 40°C. After being washed 5 timeswith distilled water, the residue was freeze-dried.

The cell wall fraction was suspended and incubated with 5 mlof 100 milliunits m l 1 endoglucanase in 2.5 mM sodium acetatebuffer (pH 5.0) for 24 h at 40°C with adding few drops of toluene.After being heated at 100°C for 15 min to inactivate the enzyme,the incubation mixture was centrifuged and then supernatant wascollected. The supernatant was passed through Dowex 50Wx8and Dowex 1x8 and dried up in rotary evaporator. The driedmaterial was dissolved in 500 fi\ of distilled water and then ethanolwas added to give final concentration of 80% (v/v). After stand-ing for 1 h at room temperature, the mixture was centrifuged andthen supernatant was collected and freeze-dried.

The cell wall fraction and tamarind xyloglucan were incu-bated with 50 milliunits purified xyloglucan specific endo-1,4-/?-glucanase (xyloglucanase) obtained from Geotrichum sp. M128strain in 2.5 mM sodium acetate buffer (pH 5.5) for 4h at 40°C(Mitsuishi 1999). After heating, enzymatic hydrolysates were pre-pared as described above.

The chromatogram of enzymatic hydrolysates was run onMerck silica gel 60 TLC plate, developed by 1-butanol: aceticacid : water (3 : 3 : 2) for 15 h or acetonitrile : water (17 : 3) for 2h at room temperature, and visualized by spraying with 1% (w/v)orcinol containing 50% (w/v) sulfuric acid and heating for fewminutes in an oven at 100°C.

Results

Distributions of xyloglucan andxylan in the epidermalcell wall—Immunogold labeling was applied to localize thedistribution of xyloglucan in the epidermal cell wall. Atfirst, a titration test was done to determine the optimal di-lution for the anti-xyloglucan antibody (Sone et al. 1989).The antibody solution was used at dilutions ranging from1 : 10 to 1 : 10,000 and the number of gold particles per/um2 was counted. The labeling density was saturated at a1 : 100 dilution (data not shown). From the results of thetitration test, the antibody was used at a dilution of 1 : 50.The distribution of xyloglucan epitope in the cell wall ofthe elongating and the elongated regions is shown in Figure1. The labeling density of the xyloglucan epitope was ob-tained from 30 cell walls (10 cells per segment from 3replicate segments). The labeling density in the intact cellwall of the elongating region was 84±22 per jum2 (Fig. la)and that of the elongated region was 362 ±87 per /urn2

(Fig. lb). After treatment with EDTA, the cell wall ofelongated region was labeled by the antibody (Fig. lc). Thelabeling density in the wall of EDTA-treated elongatedregion was 336 ±57 per /urn2. This value is almost identicalto that of intact cell wall. The treatment with EDTA fol-lowed by endoglucanase results in remarkable reduction ofthe labeling density in the cell wall of elongated region.(Fig. Id). The labeling density was 137±46 per/zm2 andreduced to approximately one third of intact or EDTA-

Dow



488 Xyloglucan cross-links in the cell wall

Fig. 1 Immunogold labeled sections of intact elongating (a) and elongated (b) regions, EDTA-treated (c) and EDTA-endoglu-canase-treated (d) elongated regions. Anti-xyloglucan heptasaccharides antibody labeled across the entire outer epidermal wall of bothintact elongating (a) and elongated regions (b). After treatment with EDTA, large amount of label was still observed (c). EDTA followedby endoglucanase treatment resulted in remarkable reduction of the labeling though labels were seen in the inner wall (d). Bars: 200 nm.

treated cell wall of elongated region. The remaining labelwas denned in inner layer of the cell wall. These experi-ments indicated that endoglucanase treatment is effective toremove xyloglucan from the cell wall of elongated region.

Xylan is often the major hemicellulosic components inthe secondary wall of dicotyledons. We investigated thedistribution of xylan by using immunogold labeling of an-ti-xylan antibody to elucidate whether xylan occurs in theepidermal cell wails of elongated region. Very low labelingof anti-xylan antibody was observed in the epidermal wallof elongated regions (Fig. 2). However, a dense distribu-tion of xylan epitope was observed exclusively in the cellwall of vascular bundles (data not shown).

Occurrence of cross-links—The outer wall of intactepidermal cells in an elongating region was filled withgranular materials (Fig. 3a inset), being mainly composedof pectic polysaccharides as determined by RFDE andchemical analysis (Fujino and Itoh 1998). Treatment of theepidermal wall with EDTA completely eliminated granularmaterials from the wall enabling the contour of CMFs to

be clearly visualized. CMFs with a helicoidal orientation inpolylamellate wall had a dimension of 13.5±2.0 nm/50CMFs including thickness of shadowing materials with 2

Fig. 2 Immnunogold labeled section of the cell wall of theelongated region. Anti-l,4-/?-D-xylose antibody labeled little onthe cell wall. Bar: 200 nm.

Dow




Fig. 3 Deep-etched images of intact and EDTA-treated outer cell walls of epidermal cells in the elongating (a) and elongated (b)regions. Inset of a: In the intact cell wall, the space between CMFs was completely embedded by granular substances, a: After treatmentwith EDTA, the cell wall architecture became exclusively consisted only of CMFs. Inset of b: In the intact cell wall, CMFs and theircross-links were visible (arrowheads), b: After treatment with EDTA, a large number of short fibrillar structures (some of them wereshown by arrowheads) could be clearly seen in the space between the CMFs and cross-linked with CMFs. Bars: 200 nm.

Dow




* • • . ' " „ • i

* • * % - -

4, v- . . ' ^

Fig. 4 Deep-etched images of outer cell walls of epidermal cells in the elongated region after treatment with EDTA and then SDSelectrophoresis sample buffer to remove protein (a), 4% potassium hydroxide to remove weak-bounded hemicellulose (b) and the endo-1,4-jff-glucanase to remove xyloglucan (c). a and b: A large number of cross-links (some of them were shown by arrowheads) were stillpresent in the cell wall after treatment with SDS electrophoresis sample buffer and 4% potassium hydroxide, respectively, c: Aftertreatment with the endoglucanase, cross-links remarkably reduced in the space between CMFs. Bars: 200 nm.

Dow




nm (Fig. 3a). Cross-links that interconnected CMFs were notseen in the wall of the elongating region after treatmentwith EDTA. The wall was exclusively consisted of CMFs inappearance of RFDE image. The outer wall of epidermalcells in an elongated region was also observed by RFDEbefore and after treatment with EDTA. The intact cell wallin this region contained CMFs oriented with a helicoidalpattern and short fibrils cross-linking between CMFs(Fig. 3b inset). After treatment with EDTA, the cell wallarchitecture being comprised of fibrious components wasobserved in detail (Fig. 3b). In the EDTA-treated cell wall,the dimension of CMFs was 13.6±1.9nm/50 CMFs andthat of cross-links was 7.3±2.0nm/30 cross-links. Thedimension of the cross-links was smaller than that ofCMFs. These cross-links were almost criss-crossed and in-terconnected with CMFs. When the dimension of cross-links was measured, cross-links located at the upper sur-face of replicated images were chosen to avoid confusionbetween cross-links and CMFs. CMFs situated below thesurface appeared thin due to insufficient shadowing ofplatinum. These results indicate that cell walls have differ-ent nature between elongating and elongated regions.

Characterization of the cross-links by RFDE coupledwith chemical and enzymatic treatments—To characterizethe chemical nature of the cross-links between CMFs in theEDTA-treated and elongated cells, the materials wereeither treated with SDS electrophoresis sample buffer toremove protein, with 4°7o potassium hydroxide to removeweakly-bounded hemicellulose, or with endo-l,4-/?-glu-canase to fragment xyloglucan before observation byRFDE. After treatment with SDS electrophoresis samplebuffer, the cross-links did not disappear. The dimension ofthe CMFs was 13.4±2.1 nm/50 CMFs and that of cross-links was 7.0± 1.2 nm/30 cross-links (Fig. 4a). After treat-ment with 4% potassium hydroxide, the cross-links werealso remained. The dimension of the CMFs was 12.9 ±1.5nm/50 CMFs and that of cross-links was 5.7± 1.0 nm/30cross-links (Fig. 4b). However, after incubation with en-do- 1,4-/?-glucanase, the number of cross-links decreased inthe cell wall. CMFs had a dimension of 12.9±2.0 nm/50CMFs (Fig. 4c). Although a large number of cross-linkswere removed from the slit-like pores between CMFs,lateral association of CMFs did not occur.

Analysis of enzymatic hydrolysates—Endoglucanasehydrolysates from the cell wall fraction of elongated regionwere characterized by TLC developed with solvent of 1-butanol, acetic acid and water (Fig. 5). Using well-definedand purified xyloglucanase, xyloglucan oligosaccharidesderived from the cell wall fraction (lane 2) and tamarindxyloglucan (lane 4) were examined. Spots in lane 4 wereidentical to xyloglucan oligosaccharides of heptasaccha-rides, octasaccharides and nonasaccharide, from upper tolower, respectively. Xyloglucanase hydrolysate from thecell wall fraction contained same xyloglucan oligosaccha-

Fig. 5 TLC of standard and enzymatic hydrolysates; lane 1,glucose (upper) and cellobiose (lower); lane 2, xyloglucanasehydrolysate from the cell wall fraction of elongated region; lane 3,endoglucanase hydrolysate from the cell wall fraction of elongat-ed region; lane 4, xyloglucanase hydrolysate from tamarind xylo-glucan. TLC was developed with solvent of butanol, acetic acidand water. In lane 4, each spots correspond to oligosaccharides ofheptasaccharides, octasaccharides and nonasaccharide, from up-per to lower. In lane 3, there are faint spots of glucose and cel-lobiose and clear spots of heptasaccharides, octasaccharides andsmall spot of nonasaccharide being identical to xyloglucan oligo-saccharides.

rides, thought it gave clear spots of heptasaccharides andoctasaccharides and small spot of nonasaccharides (lane 2).The endoglucanase hydrolysate from the cell wall fraction(lane 3) gave faint spots of glucose and cellobiose and spotsof oligosaccharides, which were identical to the spots ofxyloglucan oligosaccharides as shown in lane 4 and samepattern to lane 2. Therefore, these oligosaccharides occur-red by fragmentation of xyloglucan, suggesting that theendoglucanase fragmented xyloglucan in the cell wall frac-tion.

TLC developed with solvent of acetonitrile and waterwas done to separate low molecular weight sugars (Fig. 6).In endoglucanase hydrolysate (lane 2), there were spots ofglucose, cellobiose and galacturonic acid. Other neutralsugars were trace. However, most sugars still remained atbase. Endoglucanase hydrolysate was enriched in oligo-saccharides rather than mono- and disaccharides. Theseresults indicated that the endoglucanase fragmented rela-tively specific to xyloglucan in the cell wall fraction, though

Dow




t

Fig. 6 TLC of standard and endoglucanase hydrolysate; lane 1,xylose, glucose and cellobiose, upper to lower; lane 2, endoglu-canase hydrolysate from the cell wall of elongated region; lane 3,arabinose, galactose and galacturonic acid, upper to lower. TLCwas developed with solvent of acetonitrile and water. In lane 2,strong signal was seen in origin.

this enzyme digested more or less CMF.

Discussion

We have reported previously the difference in the cellwall architecture of intact epidermal cells between elon-gating and elongated regions in pea epicotyl (Fujino andItoh 1998). In the present study, we aimed to characterizethe chemical nature of the cross-links between CMFs. TheCMFs were 13.9 nm in diameter, while the cross-links were7.3 nm, a difference of almost two-fold. The cross-linkswere lost from the cell wall after treatment with strongalkali, which did not solubilize CMFs (McCann et al. 1990,Itoh and Ogawa 1993). Thus, we ruled out the possibilityof CMFs being the entity of the cross-links.

It is considered as one possibility, that the cross-linksare composed of structural proteins. The most well-studiedwall protein is extensin (Lamport 1969), which is a hy-droxyproline-rich glycoprotein and forms networks thatare structurally independent of the polysaccharide network(Mort and Lamport 1977). It accumulates in the cell wallfollowing cessation of elongation growth (Savada et al.1973). However, the cross-links are observed after treat-ment with SDS electrophoresis sample buffer. The SDS

electrophoresis sample buffer removes various kinds of cellwall protein (Corke and Roberts 1997). Whereas we did notanalyze an amount of the structural proteins in the pea cellwall, they are generally a minor component of the wall(<5%) (Cosgrove 1997). Therefore, it is unlikely that thecross-links are composed of structural proteins.

Xylan is another candidate for the component form-ing the cross-links. It contains a l,4-/?-xylan backbone witharabinose and glucuronic acid side chain units and hasan ability to bind CMFs. In monocotyledonous plants,xylan is considered to interconnect CMFs by formingcross-links (Carpita and Gibeaut 1993). However, xylan inthe epidermal cells of pea is present in a negligible amountas shown in Figure 2, while numerous cross-links are foundin the cell wall observed by RFDE. Furthermore, it isreported that xylan is extractable by a treatment of 4%potassium hydroxide (Coimbra et al. 1996). We analyzedthe 4% potassium hydroxide fraction in pea epicotyl,which was enriched in arabinose and xylose. The sugarenriched fraction is likely derived from xylan (data notshown). Our results show that the cross-links are not re-moved after the treatment with 4% potassium hydroxide.Therefore, there is little possibility that xylan is the entityof the cross-links.

Finally, xyloglucan is left as a possible candidate forthe entity of the cross-links. After treatment with endo-glucanase, the remarkable decrease of xyloglucan epitopesand cross-links in the cell wall was revealed by immunogoldlabeling and RFDE, respectively. Furthermore, the endo-glucanase hydrolysate is enriched in oligosaccharides,which is derived from fragmentation of xyloglucan asshown in Figure 5. This hydrolysate also contains glucoseand cellobiose, which are produced by enzymatic hydroly-sis of cellulose. These sugars are not present in amounts inthe endoglucanase hydrolysate. Therefore, endoglucanaseexclusively degrades and fragments xyloglucan in cell wallfraction. Taken together, our results strongly suggest thatxyloglucan is the entity of cross-links in the epidermal cellwall of pea epicotyl. This is a direct demonstration to showthat cross-links are made of xyloglucan in plant cell walls.The use of endoglucanase is effective for removing thecross-links from the cell wall without association andswelling of CMFs caused by the use of strong alkali asdescribed in previous reports (Itoh and Ogawa 1993,McCann et al. 1990). Low labeling of xyloglucan epitopesare still observed at the inner layer of the cell wall aftertreatment with endoglucanase. Xyloglucan at the innerwall might be tightly associated with CMFs because xylo-glucan is caught in a CMF as CMF synthesis takes place atthe plasma membrane (Hayashi 1989). This may result inresistance to digestion by the endoglucanase. Alternatively,it may also be considered that the enzyme simply could notpenetrate to the inner layer.

As described above, we demonstrated that cross-links

Dow




are composed of xyloglucan and the clear difference in thecell wall architecture is observed in the epidermal cells ofelongating and elongated tissues. In the wall of elongat-ing epidermal cells, no cross-links are observed betweenCMFs. In contrast, a large number of cross-links are ob-served between CMFs in the epidermal cell wall of theelongated region. On the other hand, the distribution ofxyloglucan epitopes by immunogold labeling is observedthroughout the cell wall of both regions. Baba et al. (1994)and Suzuki et al. (1998) reported that no cross-links existedin spite of the presence of xyloglucan at the surface ofCMFs observed by immunogold labeling coupled withnegative staining and RFDE, respectively. Therefore, theabsence of cross-links in the wall of elongating epidermalcells does not conflict with previous reports. Cross-links inthe intact cell wall of elongating epidermal cells in mungbean hypocotyl have been observed by Satiat-Jeunemaitreet al. (1992). However, we could not confirm the occur-rence of cross-links in the intact cell wall of the elongatingepidermal cells in pea epicotyl because the space betweenCMFs was completely occupied by granular pectic materi-als. Nor could we confirm the existence of the cross-linkseven after removal of the granular materials by the treat-ment with EDTA (Fujino and Itoh 1998). It is also notedthat the same result with pea epicotyl was obtained in thecell wall of elongating epidermal cell in Azuki bean (Vignaangularis) epicotyl (data not shown). We can not fully ac-count for the disparate description of the cell wall archi-tecture in the different species. The characteristic of cellwall architecture may vary depending on the tissues be-tween hypocotyl and epicotyl.

There is question how xyloglucan molecules are or-ganized in the cross-links at an ultrastructural level. Theresolution of the shadowing technique has been estimatedto be 1 to 2 nm (Willison and Rowe 1980). The dimensionof a single xyloglucan molecule is calculated to be 1.1 nmfrom its molecular conformation (Ogawa et al. 1990).Therefore, a single strand of a xyloglucan molecule couldnot be visualized by RFDE. The cross-links visualized byRFDE had a dimension of 5.7 to 7.6 nm (average 6.7 nm)including the thickness of the shadowing material. Sincethe deposition of shadowing material is kept to 2 nm inthickness, the actual dimension of the cross-links is esti-mated to be 4.7 nm. This value corresponds to that of 4 or5 xyloglucan molecules. The length of xyloglucan is esti-mated to be up to 400 nm (Vincken et al. 1997). Becausethe length is much longer than the distance between twoadjacent CMFs (McCann and Roberts 1991, McCann et al.1993), cross-links must be assembled with various confor-mation such as twisted helix and flat ribbon. In the mean-time, the dimension of CMFs observed from various kindsof specimens by RFDE are shown to be almost identical(Fig. 3, 4). Although the surface of CMFs is considered tobe coated by a monolayer of xyloglucan (Hayashi et al.

1987), we propose that this coat has no effect on the di-ameter of CMF in the shadowing image. The dimension ofCMFs measured in RFDE images have a standard devia-tion of approximately 2 nm. As the dimension of xyloglu-can molecules is 1.1 nm, the effect of the coating of xylo-glucan on the surface of CMFs is estimated to be within therange of standard deviation. This notion is also supportedby the following report. Whitney et al. (1995) who foundthat cross-links between cellulose ribbons produced byAcetobacter were formed in the presence of xyloglucan inculture medium. They also reported that the width of cel-lulose ribbons does not change in the presence or absenceof xyloglucan. However, we have no clear information toexplain how xyloglucan molecules are packed into cross-links.

The present data imply that CMFs can slip underthe strain of turgor pressure in the absence of cross-links,resulting in cell elongation. CMFs are interconnected andimmobilized by the formation of cross-links right beforethe cessation of cell elongation, resulting in the rigidifica-tion of the cell wall. Therefore, we hypothesize that theformation of cross-links between CMFs is responsible forthe cessation of cell elongation. Two proteins that modifythe cellulose-xyloglucan framework have been proposed tocontrol cell elongation. Expansins which are not recog-nized as enzymes, are believed to disrupt the hydrogenbonding between CMFs and xyloglucan (Cosgrove 1997).The cell wall of elongating cell shows high susceptibility tothe expansin action (Cho and Kende 1997, McQueen-Mas-on 1995). Endoxyloglucan transferase hydrolyzes a portionof the donor xyloglucan and transfers the fragment to thenon-reducing end of an acceptor xyloglucan (Fry 1995,Nishitani 1997). Its activity is high at a short distance backfrom the apical bud in the pea third internode (Fry etal. 1992), which corresponds to our "elongated region"(Fujino and Itoh 1998). We do not know yet the mechan-ism involved in the cross-link formation. However, eitherthe rearrangement of the cellulose-xyloglucan frameworkcatalyzed by this enzyme and protein, or self assembly ofxyloglucan for making cross-links as shown by Whitney etal. (1995) may be responsible for the formation of thecross-links in the cell wall of pea.

We address that cross-links are composed of xyloglu-can by testing with RFDE in the elongated cell wall beforeand after treatment with endoglucanase. Cell wall archi-tecture showed a difference depending on developmentalregion in the pea epicotyl. Our results directly support thehypothesis that the mechanical property of epidermal cellwall is controlled by xyloglucan cross-links between CMFs(Taiz 1984). The present investigation gives first account toshow that application of RFDE disclose dynamic changesof the molecular architecture of the cell wall in the different'stages of cell elongation.

Dow




The authors wish to thank Dr. T. Hayashi (Wood ResearchInstitute, Kyoto University) for gift of tamarind xyloglucan andvaluable discussions. We also thank Dr. S. Kimura and Mr. D.Karlson for critical reading of the manuscript. This work is sup-ported by Grant-in-Aid for "Research for the Future" Program(nos. JSPS-RFTF 96L00605) from the Ministry of Education,Science and Culture of Japan.

Reference

Baba, K., Sone, Y., Misaki, A. and Hayashi, T. (1994) Localization ofxyloglucan in the macromolecular complex of xyloglucan and cellulosemicrofibrils in pea stems. Plant Cell Physiol. 35: 439-444.

Carpita, N. and Gibeaut, D.M. (1993) Structural models of primary cellwalls in flowering plants: consistency of molecular structure with thephysical properties of the walls during growth. Plant J. 3: 1-30.

Cho, H-T. and Kende, H. (1997) Expansins and internodal growth ofdeep water rice. Plant Physiol. 113: 1145-1151.

Coimbra, M.A., Delgadillo, I., Waldron, K.W. and Selvendran, R.R.(1996) Isolation and analysis of cell wall polymers from olive pulp. InModern Methods of Plant Analysis, Vol. 17. Plant Cell Wall Analysis.Edited by Linskens, H.F. and Jackson, J.F. pp. 19-4. Springer-Verlag,Berlin.

Corke, F.M.K. and Roberts, K. (1997) Large changes in the population ofcell wall proteins accompany the shift to cell elongation. J. Exp. Bot. 48:971-997.

Cosgrove, J.D. (1997) Assembly and enlargement of the primary cell wallin plants. Annu. Rev. Cell Biol. 13: 171-201.

Fry, S.C. (1986) Cross-linking of matrix polymers in the growing cell wallsof angiosperms. Annu. Rev. Plant. Physiol. 37: 165-186.

Fry, S.C. (1995) Polysaccharide-modifying enzymes in the plant cell wall.Annu. Rev. Plant Physiol. Mol. Biol. 46: 497-520.

Fry, S.C, Smith, R.C., Renwick, K.F., Martin, D.J., Hodge, S.K. andMatthews, K.J. (1992) Xyloglucan endotransglycosylase, a new wall-loosening enzyme activity from plants. Biochem. J. 282: 821-828.

Fujino, T. and Itoh, T. (1994) Architecture of the cell wall of green alga,Oocystis apiculata. Protoplasma 180: 39-48.

Fujino, T. and Itoh, T. (1998) Change in pectin structure during epidermalcell elongation in pea {Pisum sativum) and its implications for cell wallarchitecture. Plant Cell Physiol. 39: 1315-1323.

Hahn, M.G., Bucheli, P., Cervone, F., Doares, S.H., O'Neil, R.A.,Darvill, A. and Albersheim, P. (1989) Roles of cell wall constituents inplant-pathogen interactions. In Plant-Microbe Interactions. Molecularand Genetic Perspectives, Vol. 3. Edited by Kosuge, T. and Nesdter,E.W. pp. 131-181. McGraw Hill, New York.

Hayashi, T. (1989) Xyloglucans in the primary cell wall. Annu. Rev. PlantPhysiol. Mol. Biol. 40: 139-168.

Hayashi, T. (1991) Biochemistry of xyloglucan in regulating cell elongationand expansion. In The Cytoskeletal Basis of Plant Growth and Form.Edited by Lloyd, C.W. pp. 131-144. Academic Press, London.

Hayashi, T. and Maclachlan, G. (1984) Pea xyloglucan and cellulose. I.Macromolecular organization. Plant Physiol. 75: 596-604.

Hayashi, T., Marsden, M.P.F. and Delmer, D.P. (1987) Pea xyloglucanand cellulose. V. Xyloglucan-cellulose interaction in vitro and in vivo.Plant Physiol. 83: 384-389.

Hoson, T. (1993) Regulation of polysaccharide breakdown during auxin-induced cell wall loosening. J. Plant Res. 106: 369-381.

Itoh, T. and Ogawa, T. (1993) Molecular architecture of the cell wall ofpoplar cells in suspension culture, as revealed by rapid-freezing anddeep-etching techniques. Plant Cell Physiol. 34: 1187-1196.

Laemmli, U.K. (1970) Cleavage of structural proteins during the assemblyof the head of the bacteriophage T4. Nature 227: 680-685.

Lamport, D.T.A. (1969) Isolation and partial characterization of hy-dorxyprolin-rich glycopeptides obtained by enzymic degradation of pri-mary cell wall. Biochemistry 8: 1155-1163.

McCann, M.C. and Roberts, K. (1991) Architecture of the primary cellwall. In The Cytoskeletal Basis of Plant Growth and Form. Edited byLloyd, C.W. pp. 109-129. Academic Press, London.

McCann, M.C, Wells, B. and Roberts, K. (1990) Direct visualization ofcross-links in the primary wall. J. Cell Sci. 96: 323-334.

McCann, M.C, Wells, B. and Roberts, K. (1993) Complexity in the spatiallocalization and length distribution of plant cell-wall matrix polysac-charides. J. Microsc. 166: 123-136.

McQueen-Mason, S.J. (1995) Expansins and cell wall extension. J. Exp.Bot. 46: 1639-1650.

Mitsuishi, Y. (1999) Xyloglucan specific endo-l,4-/?-glucanase from Ge-otrichum sp. M128. In Plant Biology '99 (Abstract of the AnnualMeeting of the American Society of Plant Physiologists). 136.

Mort, A.J. and Lamport, D.T.A. (1977) Anhydrous hydrogen fluoridedeglycosylates glycoproteins. Plant Physiol. 67: 397-407.

Nishitani, K. (1997) The role of endoxyloglucan transferase in the organi-zation of plant cell walls. Annu. Rev. Cytol. 173: 157-206.

Ogawa, K., Hayashi, T. and Okamura, K. (1990) Conformational analysisof xyloglucans. Int. J. Biol. Macromol. 12: 218-222.

Roberts, K. (1990) Structure at the cell surface. Curr. Opin. Cell Biol. 2:920-928.

Satiat-Jeunemaitre, B., Martin, B. and Hawes, C. (1992) Plant cell wallarchitecture is revealed by rapid-freezing and deep-etching. Protoplasma167: 33-42.

Savada, D., Walker, F. and Chrispeels, M.J. (1973) Hydroxyproline-richcell wall protein (extensin): biosynthesis and accumulation in growingpea epicotyls. Dev. Biol. 30: 42-48.

Sone, Y., Kuramae, J., Shibata, S. and Misaki, A. (1989) Immunochemi-cal specificities of antibody to the heptasaccharide unit of plant xylo-glucan. Agric. Biol. Chem. 53: 2821-2823.

Suzuki, K., Baba, K., Itoh, T. and Sone, Y. (1998) Localization of xylo-glucan in cell wall of suspension culture of tobacco by rapid-freezing anddeep-etching techniques coupled with immunogold labelling. Plant CellPhysiol. 39: 1003-1009.

Taiz, L. (1984) Plant cell expansion: regulation of cell wall mechanicalproperties. Annu. Rev. Plant Physiol. Mol. Biol. 35: 585-657.

Talbott,*L.D. and Ray, P.M. (1992) Molecular size and separability fea-tures of pea cell wall polysaccharides: implications for models of pri-mary wall structure. Plant Physiol. 98: 357-368.

Vincken, J-P., York, W.S., Beldman, G. and Voragen, A.G.J. (1997) Twogeneral branching patterns of xyloglucan, XXXG and XXGG. PlantPhysiol. 114: 9-13.

Whitney, S.E.C., Brigham, J.E., Darke, A.H., Reid, J.S.G. and Gidfey,M.J. (1995) In vitro assembly of cellulose/xyloglucan networks: ultras-tructural and molecular aspects. Plant J. 8: 491-504.

Willison, J.H.M. and Rowe, A.J. (1980) Resolution and quantitation inreplicas and shadowed specimens. In Practical methods in electron mi-croscopy, Vol. 8. Replica, Shadowing and Freeze-Etching Techniques.Edited by Glauert, A.N. pp. 245-282. North-holland publishing com-pany Amsterdam, New York.

(Received November 1, 1999; Accepted January 31, 2000)

Dow



Documents

Characterization of Cross-Links between Cellulose Microfibrils, and