Planta (1997) 203: 289±294

A role for - in epidermal cell expansion

Lei Ding, Jian-Kang Zhu Department of Sciences, University of Arizona, Tucson, AZ 85721, USA

Received: 13 February 1997 / Accepted: 1 April 1997

Abstract. Arabinogalactan-proteins (AGPs) are abund- membrane, and intercellular spaces (Fincher ant plant that react with (b-D-Glc)3 but et al. 1983; Komalavilas et al. 1991; Serpe and Nothnagel not (b-D-Man)3 . We report here that 1995). The moiety of AGPs consists of treatment with (b-D-Glc)3 Yariv reagent caused inhibi- mainly and with minor amounts of tion of root growth of (L.) Heynh. other sugars including uronic acids (Fincher et al. 1983; seedlings. Moreover, the treated exhibited numer- Komalavilas et al. 1991). The moieties of AGPs ous bulging epidermal cells. Treatment with (b-D-Man)3 are typically rich in , , , Yariv reagent did not have any such e€ects. These and glycine (Fincher et al. 1983; Showalter results indicate a role for AGPs in root growth and and Varner 1989). The primary structures of several control of epidermal cell expansion. Because treatment AGP core proteins have recently been elucidated via with (b-D-Glc)3 Yariv reagent phenocopies the reb1 (root gene cloning (Chen et al. 1994; Du et al. 1994; Cheung epidermal cell bulging) mutant of Arabidopsis, AGPs et al. 1995; Mau et al. 1995; Du et al. 1996). were extracted from the reb1-1 mutant and compared The expression of AGPs is highly regulated during with those of the wild type. The reb1-1 roots contained (Knox et al. 1989; Pennell and an approximately 30% lower level of AGPs than the Roberts 1990; Stacey et al. 1990; Schindler et al. 1995). wild type. More importantly, while the pro®le of AGPs Experiments with monoclonal antibodies to particular from wild-type roots showed two major peaks upon AGP have demonstrated that the expression of crossed electrophoresis, the pro®le of AGPs from reb1-1 AGPs correlates with cell di€erentiation (Pennell and roots exhibited only one of the major peaks. Therefore, Roberts 1990; Knox et al. 1991). For example, the AGP the reb1 phenotype appears to be a result of defective or recognized by the missing root AGPs. Taken together, this pharmacolog- MAC207 is present in all cells of vegetative meristems, ical and genetic evidence strongly indicates a function of primordia and organs, and throughout undeveloped AGPs in the control of root epidermal cell expansion. ¯ower buds, but not in speci®c cells of developing stamens and carpels (Pennell and Roberts 1990). There- Key words: Arabidopsis (reb1 mutant) ± Arabinogalactan- fore, AGPs have been proposed to serve as speci®c cell protein ± Cell expansion ± Mutant (reb1) ± root epidermal surface markers. cell bulging ± Root growth Several functions of AGPs have been established recently. Speci®c AGPs have been found to be important for of L., because addition of AGPs extracted from carrot to a two- year-old, non-embryogenic cell line resulted in the re- Introduction induction of embryogenic potential (Kreuger and van Holst 1993). Serpe and Nothnagel (1994) observed that Arabinogalactan-proteins (AGPs) are a family of treatment of rose cell suspension with ( -D-Glc) Yariv proteoglycans that are widely distributed throughout b 3 reagent, a chromophoric molecule that selectively binds the plant kingdom (Fincher et al. 1983; Basile and Basile AGPs, caused inhibition of cell proliferation. More 1987). They are found predominantly in the plasma recently, secreted AGPs in the tobacco style have been demonstrated to stimulate and guide tube growth Abbreviations: AGP = ; Gal = galactose; (Cheung et al. 1995; Wu et al. 1995). Glc = ; Man = It has long been known that AGPs bind to Yariv Correspondence to: J.K. Zhu; E-mail: [email protected]; reagents, synthetic molecules containing phenyl-b-gly- Fax: 1 (520) 621 7186 cosides (Yariv et al. 1962; Clarke et al. 1978). This 290 L. Ding and J.-K. Zhu: Arabinogalactan-proteins and root cell expansion property has been used extensively in the detection, 90 mM boric acid and 2 mM EDTA, pH 8.3). The running bu€er quanti®cation and isolation of AGPs from plant tissues was the same as the gel bu€er. The second dimension was run with and cultured cells (Clarke et al. 1978; van Holst and 1% agarose gel containing 15 lM(b-D-Glc)3 Yariv reagent. The gel and running bu€er consisted of 25 mM Tris and 200 mM glycine, Clarke 1986; Komalavilas et al. 1991; Zhu et al. 1993). pH 8.4. After completion of electrophoresis, gels were washed Because the glycosidic linkage in Yariv reagents is overnight with 2% (w/v) NaCl and dried onto ®lter paper. required to be in the b-anomeric conjugation to react with AGPs, AGPs sometimes have also been referred to as b-lectins (Jermyn and Yeow 1975; Clarke et al. 1978). Results Yariv phenylglycosides having sugars with cis-hydroxyl or other substitutions on C2 of the b-D-glycopyranosyl b-glucosyl Yariv reagent treatment of Arabidopsis pheno- determinant also do not bind to AGPs (Nothnagel and copies the reb1 root epidermal cell bulging mutant. Lyon 1986). For example, while (b-D-Glc)3 Yariv Because (b-D-Glc)3 Yariv reagent can inhibit cell growth reagent can bind AGPs, (b-D-Man)3 or (a-D-Gal)3 in rose suspension cultures (Serpe and Nothnagel 1994), Yariv reagents do not react with AGPs. These latter we were interested in determining whether it may also Yariv reagents can serve as valuable controls to distin- alter the growth of Arabidopsis thaliana seedlings, with guish AGP-related e€ects from non-speci®c e€ects when the aim of exploring a molecular genetic approach to the using Yariv reagents in biological systems. study of AGP function. When Arabidopsis seeds were In the course of initiating a genetic screen for possible sown onto media containing up to 50 lM b-glucosyl Arabidopsis thaliana mutants with an altered response to Yariv reagent, their germination was not a€ected. Yariv reagent, we made a fortuitous observation that However, subsequent root growth was substantially exposure to (b-D-Glc)3 Yariv reagent causes the bulging inhibited by as low as 10 lM b-glucosyl Yariv reagent. of root epidermal cells. The morphological alterations Figure 1A illustrates this inhibition on roots grown on resemble the phenotype of the reb (root epidermal cell an agar surface when the plates were placed vertically. bulging) mutant of Arabidopsis isolated by Baskin et al. For comparison, no root inhibition was observed on (1992). We then found that the root of the reb1-1 mutant seedlings grown on media containing (b-D-Man)3 Yariv contains AGPs that are electrophoretically di€erent reagent, which does not bind to AGPs (Fig. 1A). To from those of the wild type. These results strongly evaluate how fast the root inhibiton could occur, 4-d-old indicate a role of AGPs in controlling root epidermal cell seedlings grown on vertical agar plates without Yariv expansion. reagent were transferred onto a medium containing 10 lM(b-D-Glc)3 Yariv reagent, or as controls onto medium containing (b-D-Man)3 Yariv reagent (Fig. 1B) Materials and methods or to medium without Yariv reagent (not shown). While root growth was visible less than 1 h after the transfer on Treatment with Yariv reagents. Wild-type Arabidopsis thaliana (L.) seedlings transferred to control medium without Yariv Heynh. (ecotype Columbia) seeds were surface-sterilized and sown reagent, or with (b-D-Man)3 Yariv reagent, seedlings onto agar media containing Murashige and Skoog (Murashige and transferred to (b-D-Glc)3 Yariv reagent failed to exhibit Skoog 1962) salts, 3% sucrose, 1.2% agar, pH 5.7 and supple- appreciable root growth. Figure 1B shows that 2 d after mented with various amounts of Yariv reagents. Yariv reagents the seedling transfer, the roots on ( -D-Glc) Yariv were prepared according to the method of Yariv et al. (1962). b 3 were grown at 23 ‹ 2 °C under constant white light reagent had only a slight growth compared with 2 1 (approx. 70 lmol photons á m) á s) ). Roots were observed and substantial growth on (b-D-Man)3 Yariv reagent. The photographed directly inside the agar media using an inverted roots transplanted on (b-D-Glc)3 Yariv reagent also microscope. showed a series of sharp twists (Fig. 1B). Shoot growth of the seedlings on (b-D-Glc)3 Yariv reagent was not Extraction and electrophoretic analysis of AGPs. Wild type and inhibited until much later, which was probably an reb1-1 mutant seedlings were grown for 10 d on vertical agar plates. indirect consequence of the reduced root growth. Roots were separated from the shoots at the base of hypocotyl with A close scrutiny of roots grown inside agar medium a razor blade and collected and frozen immediately in liquid nitrogen. To extract AGPs, 5 g of frozen root or shoot samples was containing 10 lM(b-D-Glc)3 Yariv reagent revealed ground in 5 ml of extraction bu€er containing 50 mM Tris-HCl surprising morphological changes. While root epidermal (pH 8.0), 10 mM EDTA, 2 mM Na2SO5, and 1% (v/v) Triton X- cells including root hairs and non-hair cells are elongat- 100. The extraction procedure is essentially as described in van ed in untreated plants, we found that many of them Holst and Clarke (1986) except that the desalting step was omitted. became round cells when plants were grown on (b-D- Brie¯y, the homogenate was vortexed for 10 min, incubated at 4 °C Glc) Yariv reagent-containing media (Fig. 2). Normal- for 2.5 h and then centrifuged at 14 000 g for 10 min. The 3 supernatant was mixed with ®ve volumes of ethanol and the ly, roots grown inside agar media develop no or very few mixture was incubated overnight at 4 °C. The precipitate was root hairs (Fig. 2A). However, roots grown inside agar collected by centrifugation at 14 000 g for 10 min and then media which contain (b-D-Glc)3 Yariv reagent show resuspended in 50 mM Tris-HCl (pH 8.0). The ethanol precipita- numerous bulging cells (Fig. 2B,C). Thus, the epidermal tion step was repeated once before the sample was analyzed by cell bulging caused by b-glucosyl Yariv reagent is rocket and crossed gel electrophoresis. Rocket gel electrophoresis probably not restricted to hair cells. The epidermal cell was run with 1% agarose containing 15 lM(b-D-Glc)3 Yariv reagent. The gel and running bu€er consisted of 25 mM Tris and bulging occurred primarily in the elongation zone 200 mM glycine, pH 8.4. For crossed electrophoresis, the ®rst (Fig. 2B) and in developed regions of the roots dimension was run with 1% agarose in TBE bu€er (90 mM Tris, (Fig. 2C). The root cap and meristematic region L. Ding and J.-K. Zhu: Arabinogalactan-proteins and root cell expansion 291

Fig. 1A,B. (b-D-Glc)3 but not (b-D-Man)3 Yariv reagent in- hibits Arabidopsis root growth. A Arabidopsis seeds were ger- minated and grown for 7 d on medium containing 10 lM (b-D-Glc)3 or (b-D-Man)3 Yariv reagent. B Four-day-old seedlings grown on Murashige and Skoog (1962) medium were transferred onto medium containing 10 lM(b-D-Glc)3 or (b-D-Man)3 Yariv reagent for 2 d appeared not to be a€ected (Fig. 2B). Examination of vertical agar plates and harvested. Total cellular AGPs cross-sections of the (b-D-Glc)3 Yariv reagent-treated were extracted from the roots and shoots, and analyzed roots indicated that only the epidermal layer of cells by rocket gel electrophoresis. Arabinogalactan-proteins underwent abnormal radial expansion (data not shown). were detected in extracts from both the roots and shoots The Yariv reagent-treatment appears to phenocopy of reb1-1 plants (Fig. 3). However, the total level of the reb1 mutants which exhibit root epidermal cell AGPs in reb1-1 roots was only about 70% of that in bulging (Fig. 2D) and reduced root growth (Baskin et al. wild-type roots (Fig. 3). In contrast, reb1-1 shoots did 1992). Two alleles of reb1 are known; reb1-1 has bulging not contain less AGPs than the wild type (Fig. 3). root epidermal cells constitutively while reb1-2 only The extracted AGPs were subsequently analyzed by exhibits cell bulging when roots are grown at elevated crossed gel electrophoresis. The AGPs from wild-type temperatures (Baskin et al. 1992). In addition to roots could be resolved into two major peaks, while only epidermal cell bulging, roots exposed to (b-D-Glc)3 one peak was detected in reb1-1 root AGPs (Fig. 4). Yariv reagent exhibited sharp twists (Fig. 1B). The Therefore, reb1-1 roots contain altered AGPs. When twisted growth resembles the temperature-sensitive reb1 AGPs extracted from the shoots were subjected to the allele (i.e. reb1-2) when it is shifted to a restrictive same crossed electrophoresis, no di€erence was detected temperature (Baskin et al. 1992). The sharp twists are between reb1-1 and the wild type (data not shown). probably caused by asymmetric epidermal cell bulging (Fig. 2C). More bulging cells are present on the outside of the turn than on the inside (Fig. 2C). Roots grown Discussion inside or on media containing up to 60 lM(b-D-Man)3 Yariv reagent did not exhibit any epidermal cell bulging We have shown that (b-D-Glc)3 Yariv reagent-treated or twisted growth pattern. In fact, (b-D-Man)3 Yariv Arabidopsis seedlings exhibit reduced root growth and reagent did not have any e€ect on either the morphology inappropriate cell expansion at the root epidermis. or growth of Arabidopsis roots. Treatment with (b-D-Man)3 Yariv reagent which does not react with AGPs, did not have any of these e€ects. The reb1-1 mutant contains modi®ed arabinogalactan- This Yariv reagent speci®city points to AGPs as the proteins. The fact that (b-D-Glc)3 but not (b-D-Man)3 target of (b-D-Glc)3 Yariv reagent in causing the growth Yariv reagent could alter the morphology and growth of and morphological alterations. Because Yariv reagents Arabidopsis roots strongly indicates that these e€ects are are known to form large aggregates (Nothnagel and a result of interference with AGPs. To determine Lyon 1986) and are unable to penetrate the plasma whether the reb1 phenotype is correlated with defective membrane, their targets are thus limited to the cell AGPs, wild-type and reb1-1 plants were grown for 7 d in surface AGPs, i.e. plasma membrane and cell wall 292 L. Ding and J.-K. Zhu: Arabinogalactan-proteins and root cell expansion

Fig. 3. reb1-1 mutant Arabidopsis roots contain lower amounts of AGPs. Gum-arabic AGPs or AGPs from root and shoot extracts of Arabidopsis were run on an agarose-containing 15 lM(b-D-Glc)3 Yariv reagent. Lanes 1±4, 0.4, 0.3, 0.2 and 0.1 lg of as AGP standard, respectively; lanes 5±8, extract from 8.5 mg (fresh weight) of wild-type root, reb1-1 root, wild-type shoot and reb1-1 shoot, respectively

interaction of cell surface AGPs with cell wall macro- molecules. Arabinogalactan-proteins have been pro- posed to interact with other cell surface molecules (Roberts 1989). Although the exact mechanism of the binding between Yariv reagents and AGPs is not understood, it is generally believed that this interaction indicates the ability of AGPs to bind to certain b-linked glucans in the cell wall. Several b-linked polysaccharides are very abundant in plant cell walls. The most abundant one is cellulose, a b-1,4-glucan (Carpita and Gibeaut 1993). Structures similar to phenyl-b-glycosides (Yariv reagent) might also be present in the cell wall polysac- Fig. 2A±D. Treatment with (b-D-Glc)3 Yariv reagent induces epider- mal cell bulging in Arabidopsis roots. A wild-type root grown inside charides and phenolics, and AGPs may be able to bind agar medium without (b-D-Glc)3 Yariv reagent; B, C, wild type roots to these wall components. grown inside agar medium containing 10 lM(b-D-Glc)3 Yariv Multiple modes of action could be proposed to reagent; D reb1-1 root grown inside agar medium without (b-D-Glc)3 account for the role of AGPs in cell expansion. For Yariv reagent example, cell surface AGPs could be involved in wall biosynthesis or assembly. Loss of these AGPs or interference with their function by Yariv reagents may AGPs. These results clearly indicate a role of cell surface AGPs in root growth and control of root epidermal cell expansion. The conclusion is further supported by the ®nding that the root epidermal cell bulging mutant, reb1-1, contains defective AGPs. Taken together, our results strongly suggest a function of cell surface AGPs in the control of cell expansion and root growth. The inhibition of Arabidopsis root growth by (b-D- Glc)3 Yariv reagent was not unexpected because this compound has already been shown to inhibit the growth of cultured rose cells (Serpe and Nothnagel 1994). However, our observation of altered root epidermal cell morphology was surprising. During the preparation of this manuscript, Willats et al. (1996) reported similar growth and morphological e€ects of (b-D-Glc)3 Yariv reagent on Arabidopsis plant roots. However, these authors did not examine AGPs in the reb1-1 mutant. The root-inhibition phenotype could be a convenient Fig 4. Arabinogalactan-protein pro®les of wild-type and reb1-1 Arabidopsis roots as revealed by crossed gel electrophoresis. The assay to select Arabidopsis mutants which do not AGPs were extracted from 280 mg (fresh weight) each of wild-type respond to (b-D-Glc)3 Yariv reagent treatment. This and reb1-1 roots. The agarose gel in the second dimension contained type of mutant would be valuable for elucidating the 15 lM(b-D-Glc)3 Yariv reagent L. Ding and J.-K. Zhu: Arabinogalactan-proteins and root cell expansion 293 disrupt or weaken cell wall structure which is necessary Cheung AY, Wang H, Wu H-M (1995) A ¯oral transmitting tissue- for controlled cell expansion. Plasma-membrane AGPs speci®c attracts pollen tubes and stimulates their have also been proposed to interact with the actin growth. Cell 82: 383±393 Clarke AE, Gleeson PA, Jermyn MA, Knox RB (1978) Charac- cytoskeleton (Roberts 1989). Defects in AGPs could terization and localization of b-lectins in lower and higher thus lead to a disorganized actin network and subse- plants. Aust J Plant Physiol 5: 707±722 quent loss of directional cell expansion. Alternatively, Du H, Simpson RJ, Moritz RL, Clarke AE, Bacic A (1994) Yariv reagent binding could activate a signal-transduc- Isolation of the protein backbone of an arabinogalactan-protein tion cascade through cell surface AGPs and their from styles of Nicotiana alata and characterization of a interacting cellular components, and indirectly lead to corresponding cDNA. Plant Cell 6: 1643±1653 Du H, Simpson RJ, Clarke AE, Bacic A (1996) Molecular growth inhibition and other responses. characterization of a -speci®c gene encoding an arab- The results in Fig. 4 show that there is apparent loss inogalactan-protein (AGP) from Nicotiana alata. Plant J 9: 313± of a subgroup of AGPs in reb1-1 roots. That subgroup is 323 represented by the smaller peak upon crossed electro- Fincher GB, Stone BA, Clarke AE (1983) Arabinogalactan phoresis (Fig. 4). Cell surface AGPs are known to be proteins: structure, biosynthesis and function. Annu Rev Plant extremely heterogeneous in size and charge (Komalavilas Physiol 34: 47±70 Jermyn MA, Yeow YM (1975) A class of lectins present in the et al. 1991; Serpe and Nothnagel 1995; Serpe and tissues of plants. Aust J Plant Physiol 2: 501±531 Nothnagel 1996; Stohr et al. 1996). Crossed electropho- Knox JP, Day S, Roberts K (1989) A set of cell surface resis can only reveal major changes in AGPs. Subtle forms an early marker of cell position, but not changes in AGPs which might be of immense functional cell type, in the root apical meristem of Daucus carota L. signi®cance are dicult to detect. The apparent loss of a Development 106: 47±56 major AGP peak could correspond to the failure of root Knox JP, Linstead PJ, Peart J, Cooper C, Roberts K (1991) Developmentally regulated epitopes of cell surface arabinoga- epidermal cells to synthesize AGPs in the reb1-1 mutant. lactan proteins and their relation to root tissue pattern The REB1 gene may encode an AGP core protein, or an formation. Plant J 1: 317±326 enzyme directly involved in its . Alterna- Komalavilas P, Zhu J-K, Nothnagel E (1991) Arabinogalactan- tively, the REB1 gene product could indirectly regulate proteins from the suspension culture medium and plasma AGP biosynthesis or modi®cation such as glycosylation. membrane of rose cells. J Biol Chem 266: 15956±15965 It is also conceivable that the apparent di€erence in Kreuger M, van Holst G-J (1993) Arabinogalactan proteins are essential in somatic embryogenesis of Daucus carota L. Planta AGP pro®le between wild-type and reb1-1 roots involves 189: 243±248 some molecule to which AGPs bind, thereby altering the Mau S-L, Chen C-G, Pu Z-Y, Moritz RL, Simpson RJ, Bacic A, extractability of AGPs. Clarke AE (1995) Molecular cloning of cDNAs encoding the The molecular and cell biological basis for the reb1 protein backbones of arabinogalactan-proteins from ®ltrate of mutant phenotypes is not known. It has been suggested suspension-cultured cells of Pyrus communis and Nicotiana that a breakdown in the organization of cortical alata. Plant J 8: 269±281 Murashige T, Skoog F (1962) A revised medium for rapid growth microtubules or a cell wall component may be respon- and bioassays with tobacco tissue culture. Physiol Plant 15: sible (Baskin et al. 1992). Alternatively, the mutants 473±497 could be defective in establishing or maintaining the Nothnagel EA, Lyon JL (1986) Structural requirements for the polarity of root epidermal cell growth. Gene cloning of binding of phenylglycosides to the surface of protoplasts. Plant the REB1 locus might provide insights into the molec- Physiol 80: 91±98 ular nature of the gene product and its mode of action. Pennell RI, Roberts K (1990) Sexual development in the pea is presaged by altered expression of arabinogalactan protein. Nature 344: 547±549 We thank Dr. Tobias Baskin (Univ. of Missouri, Columbia, MO) Roberts K (1989) The plant extracellular matrix. Curr Opinion Cell for kindly providing us with the reb1 mutant seeds, and Dr. Eugene Biol 1: 1020±1027 A. Nothnagel (Univ. of California, Riverside, CA) for providing us Schindler T, Bergfeld R, Schopfer P (1995) Arabinogalactan with the Yariv reagents used in the early phase of this work. This proteins in maize coleoptiles: developmental relationship to cell work is supported by a grant from the National Science Founda- death during di€erentiation but not to extension growth. tion (IBN-9507191). Plant J 7: 25±36 Serpe MD, Nothnagel EA (1994) E€ects of Yariv phenylglycosides on Rosa cell suspensions: evidence for the involvement of References arabinogalactan-proteins in cell proliferation. Planta 193: 542± 550 Baskin TI, Betzner AS, Hoggart R, Cork A, Williamson RE (1992) Serpe MD, Nothnagel EA (1995) Fractionation and structural Root morphology mutants in Arabidopsis thaliana. Aust J Plant characterization of arabinogalactan-proteins from the cell wall Physiol 19: 427±437 of rose cells. Plant Physiol 109: 1007±1016 Basile DV, Basile MR (1987) The occurrence of cell wall-associated Serpe MD, Nothnagel EA (1996) Heterogeneity of arabinogalac- arabinogalactan proteins in the Hepaticae. Bryologist 90: 401± tan-proteins on the plasma membrane of rose cells. Plant 404 Physiol 112: 1261±1271 Carpita NC, Gibeaut DM (1993) Structural models of primary cell Showalter AM, Varner JE (1989) Plant hydroxyproline-rich glyco- walls in ¯owering plants: consistency of molecular structure proteins. In: A Marcus (ed) The biochemistry of plants: a with the physical properties of the walls during growth. Plant J comprehensive treatise, vol 15, Molecular biology, Academic 3: 1±30 Press, New York, pp 485±520 Chen C-G, Pu ZY, Moritz RL, Simpson RJ, Bacic A, Clarke AE, Stacey NJ, Roberts K, Knox JP (1990) Patterns of expression of the Mau S-L (1994) Molecular cloning of a gene encoding an JIM4 arabinogalactan-protein epitope in cell cultures and arabinogalactan-protein from pear (Pyrus communis) cell sus- during somatic embryogenesis in Daucus carota L. Planta 180: pension culture. Proc Natl Acad Sci USA 91: 10305±10309 285±292 294 L. Ding and J.-K. Zhu: Arabinogalactan-proteins and root cell expansion

Stohr C, Snogerup L, Pennell RI, Kjellbom P (1996) An agarose gel Wu H-M, Wang H, Cheung AY (1995) A growth electrophoresis method for the separation of arabinogalactan stimulatory glycoprotein is deglycosylated by pollen tubes and proteins. Plant J 10: 943±948 displays a glycosylation gradient in the ¯ower. Cell 82: 395± van Holst G-J, Clarke AE (1986) Organ speci®c arabinogalactan- 403 proteins of Lycopersicon peruvianum (Mill.) demonstrated by Yariv J, Rapport MM, Graf L (1962) The interaction of glycosides cross-electrophoresis. Plant Physiol 80: 786±789 and saccharides with antibody of the corresponding phenylazo Willats WGT, Knox JP (1996) A role for arabinogalactan-proteins glycosides. Biochem J 85: 383±388 in plant cell expansion: evidence from studies on the interaction Zhu J-K, Bressan RA, Hasegawa PM (1993) Loss of arabinoga- of b-glucosyl Yariv reagent with seedlings of Arabidopsis lactan-proteins from the plasma membrane of NaCl-adapted thaliana. Plant J 9: 919±925 tobacco cells. Planta 190: 221±226