Arabinogalactan protein 31 (AGP31), a putative network-forming protein in Arabidopsis thaliana cell walls?

Abstract

Background and aims: Arabinogalactan protein 31 (AGP31) is a remarkable plant cell-wall protein displaying a multi-domain organization unique in Arabidopsis thaliana: it comprises a predicted signal peptide (SP), a short AGP domain of seven amino acids, a His-stretch, a Pro-rich domain and a PAC (PRP-AGP containing Cys) domain. AGP31 displays different O-glycosylation patterns with arabinogalactans on the AGP domain and Hyp-O-Gal/Ara-rich motifs on the Pro-rich domain. AGP31 has been identified as an abundant protein in cell walls of etiolated hypocotyls, but its function has not been investigated thus far. Literature data suggest that AGP31 may interact with cell-wall components. The purpose of the present study was to identify AGP31 partners to gain new insight into its function in cell walls.

Methods: Nitrocellulose membranes were prepared by spotting different polysaccharides, which were either obtained commercially or extracted from cell walls of Arabidopsis thaliana and Brachypodium distachyon. After validation of the arrays, in vitro interaction assays were carried out by probing the membranes with purified native AGP31 or recombinant PAC-V5-6xHis. In addition, dynamic light scattering (DLS) analyses were carried out on an AGP31 purified fraction.

Key results: It was demonstrated that AGP31 interacts through its PAC domain with galactans that are branches of rhamnogalacturonan I. This is the first experimental evidence that a PAC domain, also found as an entire protein or a domain of AGP31 homologues, can bind carbohydrates. AGP31 was also found to bind methylesterified polygalacturonic acid, possibly through its His-stretch. Finally, AGP31 was able to interact with itself in vitro through its PAC domain. DLS data showed that AGP31 forms aggregates in solution, corroborating the hypothesis of an auto-assembly.

Conclusions: These results allow the proposal of a model of interactions of AGP31 with different cell-wall components, in which AGP31 participates in complex supra-molecular scaffolds. Such scaffolds could contribute to the strengthening of cell walls of quickly growing organs such as etiolated hypocotyls.

Keywords: AGP31; Arabidopsis thaliana; Brachypodium distachyon; PAC domain; arabinogalactan protein 31; glycoprotein; non-covalent interactions; plant cell wall; polysaccharides.

Fig. 1.

In vitro interactions between AGP31 and cell-wall polysaccharides. (A) One microlitre (0·5 μL for cellulose) of each polysaccharide-enriched fraction (pectins, hemicelluloses and cellulose) was spotted on a nitrocellulose membrane. Cell-wall polysaccharide-enriched fractions from A. thaliana etiolated hypocotyls or B. distachyon leaves were used for interaction with purified native AGP31 at a final concentration of 10 μg mL^–1. PNA coupled to digoxigenin was used to detect the interactions. Control was performed with PNA coupled to digoxigenin alone. (B) The same experiment was done with the hemicellulose-enriched cell-wall fraction (1 μL spotted on the membrane) treated with proteinase K at 10, 1 or 0·01 μg μL^–1. Negative controls were performed with PNA coupled to digoxigenin alone and/or no proteinase K treatment. (C) The activity of the proteinase K was checked on MBP in the NaOH solution used to extract hemicelluloses from cell walls. MBP was detected by specific antibodies. Interaction assays with AGP31 were carried out in duplicate.

Fig. 2.

In vitro interactions between AGP31 and commercial polysaccharides. One microlitre (0·5 μL for cellulose) of each polysaccharide solution (1 mg mL^–1 except RGI and apple m.e.PGA, which were also deposited at 10 mg mL^–1) was spotted on a nitrocellulose membrane: citrus and orange PGA; apple and citrus m.e.PGA; galactan; RGI; AG; XG; β-glucan; birch and beech wood xylans; AX; and cellulose. The membrane was incubated in the presence of purified native AGP31 at a final concentration of 10 μg mL^–1. Interactions were detected with PNA coupled to digoxigenin. The negative control was performed with PNA coupled to digoxigenin alone. A positive control was performed with AGP31 (1 μg) directly deposited onto the nitrocellulose membrane. Interaction assays with AGP31 were carried out in duplicate.

Fig. 3.

Purification of the recombinant PAC domain produced in N. benthamiana. The PAC-V5-6xHis recombinant protein was produced in N. benthamiana leaves during 2 d after infiltration with transformed A. tumefaciens. Proteins were separated by SDS-PAGE and stained with Coomassie blue after CEC (lane 1) and CEC + NAC (lane 4). The same fractions were analysed by western blot using V5 antibodies after CEC (lane 2) and CEC + NAC (lane 5), or by GNA lectin blot after CEC (lane 3). Molecular masses of markers (M) are in kDa. Arrows point to proteins of interest.

Fig. 4.

In vitro interactions between the recombinant PAC-V5-6xHis domain and cell-wall polysaccharides. (A) One microlitre (0·5 μL for cellulose) of each cell-wall polysaccharide-enriched fraction (pectins, hemicelluloses and cellulose) was deposited on a nitrocellulose membrane. Cell-wall polysaccharide-enriched fractions were from A. thaliana etiolated hypocotyls or B. distachyon leaves. Negative control was performed with anti-V5 antibodies alone. (B) One microlitre (0·5 μL for cellulose) of each of the following commercial polysaccharide solutions (1 mg mL^–1 except RGI, which was deposited at 10 mg mL^–1) was spotted on a nitrocellulose membrane: citrus and orange PGA; apple and citrus m.e.PGA; galactan; RGI; AG; XG; β-glucan; birch and beech wood xylans; AX; and cellulose. (C) One microlitre of the A. thaliana cell-wall hemicellulose-enriched fraction, 1 or 10 mg mL^–1 of commercial galactan or RGI were spotted on a nitrocellulose membrane. Incubation was performed in the presence of 10 mm β-mercaptoethanol. The membranes were incubated in the presence of the purified recombinant PAC-V5-6xHis domain at a final concentration of 30 μg mL^–1. Interactions were detected with anti-V5 antibodies. Interaction assays were carried out in duplicate in A and B.

Fig. 5.

Competition assays to test the specificity of the interactions between native AGP31 or PAC-V5-6xHis and polysaccharides in vitro. One microlitre of the A. thaliana hemicellulose-enriched cell-wall fraction and 1 μL of each of the following commercial polysaccharides were spotted on the nitrocellulose membrane: galactan (1 or 10 mg mL^–1); RGI (1 or 10 mg mL^–1). (A) AGP31 (10 μg mL^–1) was incubated with the membrane either directly or after pre-incubation for 1 h with galactan at 1, 5 or 10 mg mL^–1, or with galactose at 1, 10 or 100 mg mL^–1. Interactions were detected with PNA coupled with digoxigenin. (B) PAC-V5-6xHis (10 μg mL^–1) was used directly, or after pre-incubation for 1 h with galactan at 1, 5 or 10 mg mL^–1. Interactions were detected with anti-V5 antibodies.

Fig. 6.

Interactions between AGP31 and PAC-V5-6xHis in vitro. One microgram of each purified protein was fixed on the nitrocellulose membrane. (A) AGP31 (10 μg mL^–1) was incubated with PAC-V5-6xHis spotted on the membrane. The interaction was detected with PNA coupled with digoxigenin. (B) PAC-V5-6xHis (10 μg mL^–1) was incubated with AGP31 spotted on the membrane. The interaction was detected with anti-V5 antibodies.

Fig. 7.

Dynamic light scattering of purified native AGP31 in solution. Hydrodynamic radii (nm) were estimated from DLS analysis of purified native AGP31 at 0·02 mg mL^–1. Assays were conducted directly following purification in elution buffer used for the NAC step.

Fig. 8.

Model of interactions of AGP31 with cell-wall components. AGP31 is schematized according to Hijazi et al. (2012), with arabinogalactans (AG) on its AGP domain and Hyp-O-Gal/Ara-rich motifs and Ara-oligosaccharides on its Pro-rich domain. Non-covalent interactions between AGP31 and its cell-wall partners are represented with dotted line.