A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE Running head: A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE

Arabinogalactan-proteins (AGPs) are a family of extracellular plant proteoglycans implicated in many aspects of plant growth and development including in vitro somatic embryogenesis (SE). We found that specific AGPs were produced by cotton (Gossypium hirsutum) calli undergoing SE and that when these AGPs were isolated and incorporated into tissue culture media, cotton SE was promoted. When the AGPs were partly or fully deglycosylated, SEpromoting activity was not diminished. Testing of AGPs separated by reversed-phase HPLC revealed that the SE-promoting activity resided in a hydrophobic fraction. We cloned a fulllength cDNA (GhPLA1) that encoded the protein backbone of an AGP in the active fraction. It has a chimeric structure comprising an N-terminal signal sequence, a phytocyanin-like (PL) domain, an AGP-like domain and a hydrophobic C-terminal domain. Recombinant production of GhPLA1 in tobacco (Nicotiana tabacum) cells enabled us to purify and analyse a single glycosylated AGP and to demonstrate that this chimeric AGP promotes cotton SE. Furthermore, the non-glycosylated PL domain from GhPLA1, which was bacteriallyproduced, also promoted SE, indicating that the glycosylated AGP domain was unnecessary for in vitro activity. www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 5 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE TEXT Arabinogalactan-proteins (AGPs) comprise a diverse group of plant proteoglycans (see reviews, (Fincher et al., 1993; Nothnagel, 1997; Seifert and Roberts, 2007; Ellis et al., 2010)). They are structurally complex, generally consisting of a proline-, alanine-, serineand threonine-rich protein backbone which is extensively modified, principally by hydroxylation of proline residues (to hydroxyproline, Hyp) and subsequent glycosylation through Olinkages with type II arabinogalactans (Tan et al., 2003; Shimizu et al., 2005). Many AGPs also have a C-terminal hydrophobic domain which is processed and replaced with a glycosylphosphatidylinositol (GPI) anchor, which acts to tether the molecule to the extracellular face of the plasma membrane (Schultz et al., 1998). AGPs are also defined by their ability to be bound and precipitated by the synthetic dye, β-glucosyl Yariv reagent (βGlcY) and related molecules (Yariv et al., 1967). These dyes have been useful in isolating, localizing and quantifying AGPs. AGPs are grouped into three subclasses (Schultz et al., 2002). “AGPs” have an Nterminal signal sequence, an arabinogalactosylated domain and a hydrophobic C-terminal domain. “Chimeric AGPs” contain at least one arabinogalactosylated domain and a domain with an unrelated motif while “hybrid AGPs” contain arabinogalactosylated as well as different proline/hydroxyproline-rich glycoprotein motifs. AGPs are implicated in many aspects of plant cell growth and development. Historically, it was not possible to assign roles to individual AGPs as tests were conducted with unfractionated mixtures of AGPs. More recently, individual AGPs mainly from Arabidopsis have been studied using techniques such as mutant analysis and gene knockout/silencing, providing evidence for roles of individual AGPs in cell expansion, root and seed regeneration, the coordination of vascular development, both male and female gametogenesis, the development of cotton fibres and as contributors to plant stem strength (Shi et al., 2003; van Hengel and Roberts, 2003; Acosta-Garcia and Vielle-Calzada, 2004; Motose et al., 2004; Yang et al., 2007; Levitin et al., 2008; Coimbra et al., 2009; Li et al., 2010; MacMillan et al., 2010). Conditioned media from in vitro embryogenic cultures contain factors that can promote somatic embryogenesis (SE), implying the presence of secreted signalling molecules (de Vries et al., 1988). There is evidence that secreted AGPs, which are components of conditioned media, are involved in SE. For example, SE in carrot and spruce cell cultures was promoted when AGPs from conditioned media were added exogenously (Kreuger and van www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 6 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE Holst, 1993; Egertsdotter and von Arnold, 1995). Subsequent studies showed the association of particular AGP epitopes with SE-promoting activity and the involvement of AGPs in SE for several other species (Kreuger et al., 1995; McCabe et al., 1997; Toonen et al., 1997; Chapman et al., 2000; Saare-Surminski et al., 2000; Ben Amar et al., 2007). There is also evidence that SE-promoting AGPs may be cleaved by an endochitinase (Egertsdotter and von Arnold, 1988; Domon et al., 2000; van Hengel et al., 2001; van Hengel et al., 2002) but neither the identity of the individual AGP(s) involved in promoting SE nor the mechanism of action have been established. In this study, we focussed on SE in cotton, which is a limiting step in cotton transformation, and the potential role of AGPs in this process. We show that cotton (Gossypium hirsutum ‘Coker 315’) calli undergoing somatic embryogenesis secrete an AGP fraction that promotes SE when incorporated back into the growth medium. We report the cloning and sequencing of a cDNA encoding a chimeric AGP present in this fraction and show that this molecule promotes SE. www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 7 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE RESULTS Specific AGPs are Produced by Non-Embryogenic and Embryogenic Cotton Calli Cotton cell cultures were established from hypocotyl explants grown initially in the presence of 2,4-dichlorophenoxyacetic acid (2,4-D) and kinetin, which stimulate nonembryogenic callus production. Subsequent transfers to media without exogenous phytohormones result in the eventual production of embyrogenic calli. Non-embryogenic and embryogenic calli differ morphologically with the latter being brown rather than green and more friable in texture (Fig. 1A). We extracted AGPs from both types of calli using βglucosyl Yariv reagent (β-GlcY) and found that embryogenic calli had three times the amount of total AGPs compared to non-embryogenic calli (Table I). These complex AGP mixtures were analysed by SDS-PAGE. Both samples were detected as high molecular mass smears with β-GlcY (Fig. 1B). AGPs from embryogenic calli (“embryogenic AGPs”) could also be detected with Coomassie blue whereas AGPs from non-embryogenic calli (“nonembryogenic AGPs”) could not. No co-precipitating proteins (non-AGP) were detected when AGP samples were stained with Coomassie blue. Non-embryogenic AGPs comprised a single, hydrophilic peak when analysed by reversed-phase HPLC whereas embryogenic AGPs comprised four main peaks, three of which were relatively hydrophobic (Fig. 1C). All four HPLC fractions of the embryogenic AGPs were detected with β-GlcY on SDS-PAGE but only the three hydrophobic fractions were detected with Coomassie blue (Fig. 1D). Embryogenic AGPs Promote SE To test the effect of embryogenic AGPs on cotton SE, we developed a bioassay that involved transferring cultured cotton hypocotyl explants onto tissue culture media in which AGPs had been incorporated. After 4 weeks, the number of explants that had developed embryogenic calli was then tallied. There were ten bioassays testing the effect of tissue culture media containing embryogenic AGPs (1 mg/L) on cotton SE compared to the effect of control media (no AGPs) (Table I and Table S1). When grown on control media, the percentage of lines that developed embryogenic calli four weeks after the first transfer ranged from 0% (Bioassay 1) to 43% (Bioassay 6). In the presence of embryogenic AGPs, this percentage ranged from 18% (Bioassay 3) to 73% (Bioassay 6). There was considerable trialto-trial variation in the percentage of explant lines developing embryogenic calli although the effect of embryogenic AGPs in increasing the percentage of embryogenic lines was consistent. This effect was statistically significant (p < 0.001) when we analyzed the results www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 8 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE by binary logistic regression. The odds of an explant line developing embryogenic calli on media containing embryogenic AGPs were more than twice the odds for an explant line doing so when grown on control media (Table II). Embryogenic calli that developed in the bioassay led to the production of somatic embryos that could be regenerated into fertile plants using established protocols (Umbeck et al., 1987). Embryogenic AGPs also promoted SE at media concentrations of 2 and 4 mg/L but there were no significant differences between the effects of embryogenic AGPs at the different concentrations. As described below, this AGP sample contained fractions that either promoted, inhibited or had no effect on SE. AGPs extracted from non-embryogenic calli inhibited SE when incorporated into tissue culture media and this effect was statistically significant (Table II and Table S1). The odds of an explant line producing embryogenic calli in the presence of non-embryogenic AGPs was reduced approximately 5-fold compared to growth on control media. In contrast, there was no significant effect when AGPs extracted from gum arabic were tested (Table II and Table S1). SE-Promotive AGPs are Extracellular To investigate whether the SE-promotive AGPs were extracellular, we extracted the extracellular pool of AGPs from embryogenic calli by stirring wet calli in a detergent-free buffer and then precipitating the AGPs using β-GlcY. The yield of extracellular AGPs was less than that of total AGPs (Table I) although the two AGP samples had similar RP-HPLC profiles apart from the absence of the most hydrophobic fraction (fraction 4, Fig. 1C) in the extracellular sample (data not shown). In the SE bioassay, we found that both AGP samples (2 mg/L) increased the number of explant lines developing embryogenic calli compared to the control medium. The increase was statistically greater with the extracellular AGPs (Table II), indicating that this fraction was enriched with the SE-promotive component. RP-HPLC Fraction 2 of Embryogenic AGPs has SE-Promoting Activity To identify the SE-promotive component further, we fractionated the total embryogenic AGPs by RP-HPLC into fractions 1 to 4 (Fig. 1C) which comprised approximately 75%, 4%, 11% and 10% respectively, of the total AGPs by weight. These fractions were then tested individually in the SE bioassay at concentrations that reflected their proportion of total embryogenic AGPs (Table II and Table S1). Fraction 1, when incorporated into tissue culture media at 1.5 mg/L inhibited the number of lines developing embryogenic calli, compared to the control medium. Fraction 2, incorporated into media at 0.08 mg/L, promoted SE while www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 9 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE fractions 3 and 4, incorporated at 0.22 mg/L and 0.20 mg/L, respectively, did not have significant effects. When analysed by SDS-PAGE, fractions 2 and 3 and to a lesser extent, fraction 4, stained positive for protein with Coomassie blue, with broad, high molecular weight smears (Fig. 1D). When the same gel was stained with β-GlcY, these smears stained positive for AGPs; fraction 1 stained positive for AGPs but not protein. Embryogenic AGPs that are Altered in Carbohydrate Content Retain SE-Promoting Activity To determine the importance of the carbohydrate component of the AGPs on SEpromotion, we tested embryogenic AGP samples that had been partly or fully deglycosylated using trifluoroacetic acid (TFA) or anhydrous hydrofluoric acid (HF), respectively. The neutral monosaccharide composition of the embryogenic AGPs comprised mainly galactosyl (64 mol%) and arabinosyl (30%) residues. TFA-treatment resulted in the removal of almost all of the arabinosyl residues (Fig. 2A) and a reduction in size of the Coomassie-stained smear compared to the native AGPs by SDS-PAGE (Fig. 2B). The HF-treated AGPs were further reduced in size. Neither of the treated AGPs could be detected with β-GlcY. In the SE bioassay, we found that the number of lines developing embryogenic calli increased in the presence of both acid-treated AGPs compared to the control medium and that this increase was statistically significant (Table II and Table S1). A Full-Length cDNA was Cloned, Sequenced and Named GhPLA1 To identify molecules present in HPLC fraction 2 which might be responsible for the SE-promotive activity, we attempted to clone cDNA(s) encoding protein backbone(s) of AGPs in this fraction. Three peptide sequences were obtained from a tryptic digest of this fraction as well as one sequence from the native sample. Two of the sequences (FQIGDSLV and KEIMVGGKTGAWKIP) were similar to plastocyanin-like proteins, one sequence was similar to a lipid binding protein (STASLGVTLSV), while one sequence returned few BLAST results (EDYTSXXTSNPIAEYK where X indicates that no amino acid residue was detected). Using these peptide sequences, we designed gene-specific, degenerate primers that were used to amplify and clone a full-length cDNA which we designated Gossypium hirsutum phytocyanin-like arabinogalactan-protein 1 (GhPLA1) (Fig. 3). The deduced protein sequence matched three of the four peptide sequences obtained experimentally by protein sequencing. GhPLA1 is a chimeric AGP with a 175 amino acid residue backbone www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 10 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE comprising an N-terminal signal sequence, a phytocyanin-like (PL) domain, an AGP-like domain and a hydrophobic C-terminal domain. Sequencing of the corresponding genomic DNA for GhPLA1 revealed a 78 bp intron. GhPLA1 Expressed by Tobacco BY-2 Cells has SE-Promoting Activity To test whether GhPLA1 was responsible for SE-promoting activity, we recombinantly expressed GhPLA1 in tobacco BY-2 cells to obtain a sufficient, pure amount for testing. Suspension-cultured BY-2 cells expressing GhPLA1 under the control of the CaMv 35S promoter (35S-GhPLA1 construct) secreted AGPs comprising two distinct RP-HPLC fractions, one hydrophilic and one hydrophobic (labelled a and b, respectively, Fig. 4A). Tobacco cells transformed with the empty pBIN19 vector secreted AGPs in one major RPHPLC fraction which had a similar profile and retention time to the hydrophilic fraction from the 35S-GhPLA1 line (fraction a). Four peptides from GhPLA1 were identified by mass spectrometry analysis of trypsin-derived peptides from fraction b (Fig. 4B). SDS-PAGE analysis showed staining for both protein and AGP which was similar to the staining for HPLC fraction 2 of the embryogenic AGPs (Fig. 4C). The RP-HPLC retention times of these fractions were also consistent, taking into account the 60 min run time in Fig. 1C and the 30 min run time in Fig. 4A. Both samples had similar carbohydrate contents and compositions (Table IV). Linkage analyses based on the identification of permethylated alditol acetates revealed that both samples had 1,3-, 1,6and 1,3,6-Galp branch points and terminal Araf residues with a ratio of terminal linkages to branching points of ~ 1. The hydrophilic AGP fractions from both BY-2 lines showed minimal protein staining and strong AGP staining (Fig. 4C); the carbohydrate content of the samples were similar and were higher than those of the GhPLA1-containing samples (Table IV). These AGPs were used as controls (0.5 mg/L) in SE bioassays with recombinantly-expressed GhPLA1 (rGhPLA1) tested at final concentrations of 0.05 and 0.5 mg/L. The number of lines developing embryogenic calli increased in the presence of rGhPLA1 compared to the control lines and this effect was statistically significant at both concentrations (Table III and Table S2); rGhPLA1 had greater SE-promoting activity at 0.5 mg/L than at 0.05 mg/L. There were no significant differences between either of the two AGP control lines and the corresponding lines with no added AGP (Table III and Table S2). www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 11 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE The Phytocyanin-Like Domain of GhPLA1 (PL1) has SE-Promoting Activity We hypothesised that the PL domain of GhPLA1 was responsible for SE-promoting activity. This domain (PL1) was expressed in E. coli as a fusion protein with an N-terminal hexahistidine tag and a thrombin cleavage site. Following purification and thrombin cleavage, the first eight amino acid residues were confirmed by Edman degradation while the mass of the protein determined by mass spectrometry (12,772 D) was consistent with the predicted mass (12,755 D). We then tested PL1 in the SE bioassay at 0.5 mg/L. This concentration was an excess, based on previous results with rGhPLA1, but was chosen to safeguard against issues such as incorrect folding. We found that over fourbioassays, the number of lines developing embryogenic calli increased in the presence of PL1 compared to the effect of control media and this effect was statistically significant (Table III and Table S2). In contrast, the presence of ovalbumin at 0.5 mg/L had no statistically significant effect on SE. DISCUSSION This study was aimed at enhancing SE in cotton which is a limiting step in cotton transformation. Studies in carrot, spruce and cyclamen showed that AGPs derived from embryogenic culture medium, when added back to the culture media, can promote SE (Kreuger and van Holst, 1993; Egertsdotter and von Arnold, 1995; Kreuger et al., 1995). We initially compared the AGPs produced by embryogenic and non-embryogenic cotton cell cultures. The AGPs were extracted using β-GlcY which typically results in a complex mix of molecules. The higher levels of total AGPs produced by embryogenic calli compared to nonembyrogenic calli (Table I) are consistent with the increased levels of AGPs during SE in other species (Chapman et al., 2000; Saare-Surminski et al., 2000). Many of the AGPs produced by embryogenic cotton calli are different to non-embryogenic AGPs, as shown by SDS-PAGE and RP-HPLC analyses (Fig. 1). Three hydrophobic AGP fractions specific to embryogenic callus (Fig. 1C) are likely to have sections of non-glycosylated protein backbone as they were detected with both β-GlcY and Coomassie blue (Fig. 1D). In contrast, the hydrophilic AGP fractions that partially overlap by RP-HPLC are likely to be highly glycosylated as they could not be detected with Coomassie blue but were strongly detected with β-GlcY (Fig. 1). The different nature of AGPs found in embryogenic and nonembryogenic tissue has also been observed in other systems using crossed gel electrophoresis and immunological analyses (Kreuger and van Holst, 1993; Egertsdotter and von Arnold, www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 12 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE 1995; Kreuger and van Holst, 1995; Saare-Surminski et al., 2000; Samaj et al., 2008). Specific AGPs or AGP epitopes are therefore produced as SE is initiated and/or proceeds. We established a bioassay to test the effect of AGPs on cotton SE when added to the culture medium. Embryogenic AGPs promoted SE whereas non-embryogenic AGPs were inhibitory (Table II). These results are consistent with the presence of SE-promotive and inhibitory AGPs found in carrot and tomato seeds and in the conditioned media of embryogenic suspension cultured cells (Kreuger and van Holst, 1995). In cotton, the promotive AGPs could be isolated from extracts made by stirring fresh (non-lyophilized) cells in a detergent-free buffer and are therefore extracellular. Of the four embryogenic AGP fractions separated by HPLC (Fig. 1C), only fraction 2, which comprises about 4% of the total embryogenic AGPs, promoted SE. It is likely that this fraction includes at least two protein backbones as only three of the four peptides identified in the tryptic digest were encoded in the cDNA obtained. This cDNA was designated GhPLA1 (Fig. 3). ESTs corresponding to GhPLA1 have been found in embryogenic cotton calli in two independent studies where genes expressed by embryogenic calli compared to nonembryogenic calli were differentially screened (Zeng et al., 2006; Wu et al., 2009). Tobacco BY-2 cells were used to produce recombinant GhPLA1. Protein and carbohydrate analyses were used to show that recombinant GhPLA1 is similar to the predicted/native molecule (Fig. 4, Table IV). Functional testing confirmed that this molecule is an active SE-promoting AGP (Table III and Table S2). GhPLA1 shares sequence similarity to the phytocyanins, a family of plant-specific, blue type 1 copper-binding proteins that are proposed to have roles in redox reactions in non-photosynthetic plant tissues (Nersissian et al., 1998). Phytocyanins consist of three subfamilies (stellacyanins, plantacyanins and uclacyanins) which all bind copper and a fourth subfamily comprising phytocyanins which are unlikely to, based on the absence of at least two of the four conserved copper-binding residues (Nersissian and Shipp, 2002). GhPLA1 is most similar to members of this fourth subfamily which includes phytocyanin-related early nodulin (ENOD) proteins from Medicago (Greene et al., 1998) and soybean (de Blank et al., 1993; Kouchi and Hata, 1993) as well as proteins from non-legumes (Fig. 5) (Kyo et al., 2000; Yoshizaki et al., 2000; Nersissian and Shipp, 2002; Mashiguchi et al., 2004; Khan et al., 2007; Mashiguchi et al., 2009; Ma and Zhao, 2010). In addition to sequence similarity in the phytocyanin-like domain, these proteins have an N-terminal signal sequence, a variable-length AGP-like domain which is likely to be glycosylated and a hydrophobic C-terminal domain which may be GPIwww.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 13 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE anchored. These AGPs are classified as chimeric AGPs as they consist of a phytocyanin-like (PL) domain, which is not a proline/hydroxyproline rich glycoprotein motif and a domain that comprises an AGP motif (Schultz et al., 2000). We suggest that this subgroup be called the phytocyanin-like AGPs, or PLAs. While the function(s) of PLAs are not known, roles in cell wall reorganisation and organ differentiation have been proposed (Greene et al., 1998; Yoshizaki et al., 2000). Moreover, parallels between root nodule development in legumes and SE have been found: rhizobial Nod factors rescued a carrot somatic embryo mutant and promoted SE in spruce (de Jong et al., 1993; Dyachok et al., 2002). The widespread occurrence of ENOD gene homologues in nonlegumes suggests that ENOD genes have been appropriated from other roles in plant growth and development (Reddy et al., 1999). For example, the embryogenic pollen-abundant phosphoprotein/PLA from tobacco, NtEPc, is highly and specifically expressed in cells undergoing dedifferentiation from immature pollen grains to embryogenic cells (Kyo et al., 2000; Kyo et al., 2002). The expression of NtEPc marks the acquisition of pollen embryogenic competence, though it is not known whether it actively promotes this process. AGPs have also been implicated in embryo development in microspore cultures of canola, barley, maize and wheat (Paire et al., 2003; Borderies et al., 2004; Letarte et al., 2006; Tang et al., 2006; Shim et al., 2009). Pollen or microspore embryogenesis, like SE, is an example of plant cell totipotency. The sequence similarity between GhPLA1 and NtEPc suggests that the mechanisms of these processes are similar. We hypothesize that GhPLA1 promotes SE through a signalling or messenger role. This is based on its extracellular location in embryogenic cotton calli and on the very low concentrations at which PL1 and GhPLA1 can promote SE in cotton. These concentrations are consistent with the low concentrations of AGPs applied to promote SE in carrot and spruce (Kreuger and van Holst, 1993; Egertsdotter and von Arnold, 1995; van Hengel et al., 2001). GhPLA1 may function through an interaction between its PL domain and another protein. Four highly conserved surface residues (Asp60, Tyr/Phe81, His/Tyr109 and Lys123, numbering based on GhPLA1) were noted from molecular modelling studies of the Medicago PLA, MtENOD16 and multiple sequence alignments (Greene et al., 1998). These residues form a surface patch which suggests involvement in protein-protein interactions. The prediction of a GPI anchor in many PLAs including GhPLA1 provides mechanisms by which AGPs might function as signalling molecules (Schultz et al., 2000). As GhPLA1 is most likely located in the extracellular space, as suggested by its extraction with a www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 14 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE non-detergent buffer, it is probably cleaved from the plasma membrane which allows its diffusion or transport through the cell wall. Once present in the extracellular space, it presumably acts on other cells, possibly binding to another protein in doing so. Outside the cell, the AGP domain is unnecessary for SE-promoting activity because non-glycosylated PL1 protein was active in the bioassay (Table III). A role for the AGP domain might therefore involve the passage of the PLA through the cell wall to the cell surface (Borner et al., 2002; Xu et al., 2008). For example, AGPs have been proposed to act as cell wall plasticizers, acting on the pectic network to increase wall porosity and enhance cell expansion/extension (Lamport et al., 2006). Alternatively, the glycosylated AGP domains may act as a protective shield, preventing contact between the non-AGP domain and components of the plasma membrane (Borner et al., 2002). Such roles for AGP domains are intriguing when the chimeric AGPs are considered (Knox, 2006). These AGPs contain protein domains of apparently diverse structure and function (for example, fasciclins, nonspecific lipid transfer proteins and phytocyanin-like proteins). Specific functions are likely to reside in the non-AGP domain while a more general function may be attributable to the common AGP domain, though the importance of variations in domain size and number, and the nature and extent of glycosylation, is not known. The promotive activity of bacterially-expressed PL1 (Table III) confirms that the phytocyanin domain of GhPLA1 is responsible for SE-promoting activity and that the carbohydrate component of the molecule is not required, at least in vitro. This is consistent with the retained activity of the embryogenic AGPs following both de-arabinosylation and full deglycosylation (Table II). These results contrast with other studies. Monoclonal antibodies raised against carbohydrate epitopes of AGPs have been used to show that particular AGP epitopes such as JIM8 and ZUM18 are associated with the promotion of SE (Kreuger and van Holst, 1995; McCabe et al., 1997; Toonen et al., 1997). The importance of the carbohydrate component of AGPs for the promotion of SE was also suggested from experiments involving various treatments of immature carrot seed AGPs (van Hengel et al., 2001). The contrasting results between this and the present study may be due to differences in the source of the AGPs (immature carrot seeds versus embryogenic cotton calli) and the bioassay system (suspension cultured carrot protoplasts resulting in a single-cell origin of somatic embryos versus cotton hypocotyl explants resulting in a callus stage and embryos via indirect SE). It is also possible that the mechanisms of SE in these systems are regulated by different pathways and different AGPs. www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 15 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE There is now substantial evidence that AGPs are involved in the regulation of developmental processes such as SE. In the case of cotton SE, the mechanisms by which this regulation occurs is yet to be determined. However the identification of GhPLA1 as a chimeric AGP that promotes SE is a promising lead. MATERIALS AND METHODS Plant Material and Cotton Cell Cultures Cotton (Gossypium hirsutum ‘Coker 315’) cell cultures were established as follows. Seeds were surface sterilized and then germinated on media containing 15 g/L glucose, 2.17 g/L MS Basal Salt Mixture (Phytotechnology Laboratories), 0.5 mL/L Gamborg’s 1000x Vitamin Solution (Sigma-Aldrich) and 2 g/L Gelrite (Phytotechnology Laboratories), pH 5.8 for 10 days in the dark at 28°C. Hypocotyl explants (1 to2 cm) were excised and placed on control media which contained 30 g/L glucose, 4.33 g/L MS Basal Salt Mixture, 1.9 g/L KNO3, 0.9 g/L MgCl2, 0.1 g/L myo-inositol, 1 mL/L Gamborg’s 1000x Vitamin Solution, 0.45 μM 2,4-D, 0.46 μM kinetin and 2 g/L Gelrite, pH 5.8. Explants (3 to 5 per Petri dish) were grown for 5 weeks at 28°C and a 16 h/8 h light (5 to 15 μmol m s)/dark cycle. Explants were then transferred to fresh control media without exogenous plant growth regulators every 4 weeks. From about 8 weeks, the calli could be classified as either nonembryogenic or embryogenic based on morphology (Fig. 1A). Calli were harvested for AGP extractions 12 weeks after the first transfer. Alternatively, embryogenic calli were produced without exogenous plant growth regulators in the media. In this method, small amounts of non-embryogenic calli were produced from hypocotyl explants after 5 weeks. Following transfer to fresh media and a further 4 to 8 weeks of growth, embryogenic calli developed. Extraction of AGPs For total AGPs, cotton calli was lyophilized, ground with liquid nitrogen and stirred with an extraction buffer consisting of 50 mM Tris.HCl, 10 mM EDTA, 1% (v/v) Triton X100, 0.1% (v/v) β-mercaptoethanol, pH 8 (40 mL/g of calli (dry weight)) for 3 h at 4°C. Alternatively, extracellular extracts were made by stirring fresh cells with a buffer consisting of 50 mM Tris.HCl, 10 mM EDTA, 1.5% (w/v) glucose, 0.1% (v/v) β-mercaptoethanol, pH 8 (1.25 mL/g of wet calli). For extracellular AGPs, the yield was estimated using a factor of 0.045 which is the proportion of dry embryogenic calli obtained from wet calli. AGPs were precipitated from extracts using β-glucosyl Yariv reagent (β-GlcY), which was synthesized www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 16 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE (Ganjian and Basile, 1997), and the resultant complexes were dissociated using sodium dithionite (Schultz et al., 2000). AGPs were also extracted from gum arabic (Sigma-Aldrich). AGPs were quantified by weight. SDS-PAGE analyses were performed using NuPAGE Novex Bis-Tris 4-12% gels using MES or MOPS buffer (Invitrogen). Proteins were stained with Coomassie blue (GelCode Blue, Pierce) and AGPs were stained by first soaking the gel in 1% (w/v) NaCl, incubating the gel overnight in 0.02% (w/v) β-GlcY in 1% (w/v) NaCl and destaining with 1% (w/v) NaCl. HPLC Analytical reversed-phase (RP)-HPLC was performed with a Brownlee Aquapore RP300 column (C8, 2.1 mm x 100 mm, Perkin-Elmer Applied Biosystems) using a Beckman System Gold HPLC. AGPs (~ 1 mg) were loaded onto the column which was equilibrated with 0.1% (v/v) TFA and eluted with a linear gradient to 100% solvent B (80% (v/v) acetonitrile in 0.089% (v/v) TFA) over 30 or 60 min at a flow rate of 0.5 mL/min. For semipreparative purification of AGP fractions, AGPs (5 to 10 mg) were loaded onto a Zorbax 300 SB-C8 column (9.4 mm x 25 cm; Agilent Technologies) using the same buffers with a linear gradient to 100% solvent B over 60 min at a flow rate of 3 mL/min. Separation was monitored by absorption at 215 nm and 280 nm. Deglysosylation and De-Arabinosylation of AGPs AGPs were deglycosylated with anhydrous hydrofluoric acid (HF) (Mort and Lamport, 1977). AGPs were de-arabinosylated by incubation in 0.2 M TFA (2 mg/mL) at 100°C for 2 h. The mixture was then cooled and the TFA removed by rotary evaporation. Treated AGPs were tested in the SE bioassay at a concentration based on the amount of AGP prior to treatment. Somatic embryogenesis bioassay The bioassay exploits the ability of hypocotyl explants derived from a regenerable cotton cultivar, Coker 315, to produce calli on media with no added plant growth regulators. Hypocotyl explants (1 cm to 2 cm) were grown for 5 weeks on plant growth regulator-free control media as described above. The explants and the calli produced were then transferred to fresh control media or media in which a filter-sterilized solution of AGPs was incorporated just prior to the media setting (50°C to 60°C). The explants/calli were grown for a further 4 www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 17 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE weeks and then the number of explants that had developed embryogenic calli was tallied. Typically, between 50 and 100 explant lines were tested for each AGP sample and control line for each trial. Results were analysed by binary logistic regression (MINITAB 14.1). The results of this analysis were expressed as an odds ratio which is the odds of an explant line developing embryogenic calli in the presence of one AGP sample over the odds of this occurring in the presence of no AGP (control) or another AGP sample. Therefore, an odds ratio > 1 indicates an SE-promotive effect, an odds ratio ~ 1 indicates no effect and an odds ratio < 1 indicates an SE-inhibitory effect. The 95% confidence intervals for the odds ratios (95% CI) were also calculated. If the 95% CI encompassed the value of 1, the media compositions were considered to have no significant differences in their effect on SE. The statistical significance of the results were further tested by calculating p-values which reflect the likelihood that the observed odds ratio comes from two populations where the underlying odds ratio is 1 (no difference in the effect on SE). A difference (either promotive or inhibitory) was considered significant if p < 0.05. AGP Identification and Cloning Embryogenic AGP fraction 2 (> 0.1 mg), was solubilized in 50 mM NH4HCO3, pH 7.8 (100 μL). The sample was reduced with 10 mM DTT (2.5 μL) at 55oC for 1 h and then acetylated with 10 mM iodoacetamide (10 μL) at room temperature in the dark for 1 h. A further aliquot of 10 mM DTT (2.5 μL) was added and then the mixture was digested with a 20 μg aliquot of sequencing-grade trypsin (Promega) at 37°C for 16 h. Neat formic acid (1 μL) was added to stop the reaction and the resultant peptides were purified by RP-HPLC and N-terminal sequenced by automated Edman degradation. This AGP fraction was also directly N-terminal sequenced (i.e., without prior trypsin digestion). RNA was isolated from embryogenic cotton calli using Trizol LS reagent (Gibco BRL) and used to synthesize cDNA using an oligo(dT) primer and reverse transcriptase (3’RACE System, Invitrogen). The following gene-specific, degenerate primer was used in a PCR reaction in conjunction with a reverse primer based on the sequence of the oligo(dT) primer (I = inositol): F2a, 5’-AAC/T CCI ATI GCI GAG/A TAT/C AA-3’. The resultant DNA fragment was cloned into the vector, pGEM-T EASY (Promega) and sequenced (Australian Genome Research Facility, Brisbane). Nested primers based on the partially cloned sequence were then designed: F2aO, 5’-GCT ATT TCT ATA GCA ACT CAA C-3’ and F2aI, 5’-CAA ACT CAA AAC AAC CCC AAA ACC-3’. These primers were used in PCR reactions to amplify from the 3’ end to the 5’ end of the genes in www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 18 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE conjunction with the appropriate nested primers by 5’RACE (FirstChoice RLM-RACE kit, Ambion). The resultant DNA fragment was cloned into pGEM-T EASY and sequenced. Genomic DNA was isolated from a cotton leaf (DNeasy Plant Mini Kit, Qiagen) and used as a template for PCR (Platinum Taq DNA Polymerase High Fidelity, Invitrogen). The primer pairs were designed to encompass the full-length gene and had BamH I and Sal I restriction sites at the 5’ and 3’ ends, respectively: F2af, 5’-GGA TCC ATG GCT GCT AAA GC-3’ and F2ar, 5’-GTC GAC TCA AAA CAA CCC CAA AAC C-3’. Sequence data from this article can be found in the GenBank/EMBL databanks under accession numbers DQ868985 and DQ868986. Recombinant Expression The full-length cDNA for GhPLA1 was placed under the control of the 35S cauliflower mosaic virus promoter and terminator and then subcloned into the pBIN19 plant binary vector (35S-GhPLA1 construct) (Bevan, 1984). This construct was introduced into Agrobacterium tumefaciens strain LBA4404 by electroporation which was then used to transform suspension cultured tobacco cells (Nicotiana tabacum ‘Bright Yellow 2’ (BY-2)) (Newman et al., 1993). A transformation using an empty pBIN19 vector was also performed. Transformed calli were maintained on solid media supplemented with kanamycin (0.1 mg/mL) and carbenicillin (0.5 mg/mL) and used to establish suspension cultures which were subcultured weekly (5% of culture to fresh media) and grown at 28°C with gentle shaking (150 rpm); suspension cultures were scaled up to 250 mL in 1 L flasks. Following subculturing, one week-old cultures were centrifuged (30 min, 25,000 g) and the cells washed with water (0.25 original culture volume). Pooled supernatants were lyophilized and then solubilized in water (0.2 original volume). The solution was ethanol-precipitated (4 to 5 volumes) overnight at -20°C and the resultant precipitate was solubilized in 1% (w/v) NaCl. AGPs were then purified using β-GlcY. The phytocyanin-like (PL) domain of GhPLA1 (residues 26 to 138) was recombinantly expressed in E. coli as a fusion protein with an N-terminal hexahistidine tag separated from the phytocyanin-like domain with a thrombin cleavage site. The following primer pair was used to amplify the PL1 domain with a thrombin cleavage site from the full-length cDNA: F2aThr, 5’-CAC CCT GGT TCC GCG TGG ATC CAA AGA AAT CAT GGT TGG TGG CAA AAC-3’ and F2aPcy, 5’-CTA GAT TCC AAT GTA CCT ATG CTT TTG AGA C-3’. The DNA fragment was subcloned into pENTR/D-TOPO (Invitrogen) and transferred to www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 19 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE pDEST17 by site-specific recombination (Gateway Technology, Invitrogen). The construct was expressed in E. coli (BL21 Star (DE3), Invitrogen). Proteins were extracted under denaturing conditions using immobilized metal affinity chromatography (IMAC, Ni-NTA Agarose, Qiagen). The recombinant protein was then treated with immobilized thrombin (Thrombin CleanCleave Kit, Sigma-Aldrich); the cleaved hexahistidine tag was removed by IMAC under native conditions. Mass Spectrometry MS and MS/MS data were acquired on a QSTAR XLTM hybrid quadrupole-TOF instrument (Applied Biosystems/MDS Sciex) using the AnalystQS software (Applied Biosystems/MDS Sciex) in a data-dependent acquisition mode. Spectra were then searched using MASCOT software (Matrix Science) run on an in-house server. MS/MS spectra were also validated manually. MALDI-TOF mass spectrometry was performed on a Voyager-DE STR instrument (Applied Biosystems) with sinapinic acid (Sigma-Aldrich) as the matrix. Carbohydrate Analyses Total carbohydrate content was determined by the phenol-sulfuric assay (Dubois et al., 1956) using galactose as the standard. Uronic acids were estimated using a modified sulfamate/m-hydroxydiphenyl assay (Filisetti-Cozzi and Carpita, 1991). The levels of free and esterified uronic acids were determined by carboxyl-reduction (Kim and Carpita, 1992). To determine the linkage and substitution patterns, carboxyl-reduced AGPs were methylated with CH3I (Ciucanu and Kerek, 1984). Neutral monosaccharides were determined after alditol acetate derivatization (Albersheim et al., 1967). The alditol acetates and the permethylated alditol acetates were recovered in CH2Cl2 and analysed by gas chromatography-mass spectrometry (GC-MS) (Zhu et al., 2005). Sequence Analysis Searches for similar sequences were performed using BLAST programs provided by the National Center for Biotechnology Information (Altschul et al., 1990). Alignments were performed using ClustalW (Thompson et al., 1994). The N-terminal signal sequence was predicted using SignalP (Bendtsen et al., 2004) while the presence of a GPI-anchor was predicted using the big-PI Plant Predictor (Eisenhaber et al., 2003). www.plantphysiol.org on July 16, 2017 Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. 20 A CHIMERIC AGP PROMOTES SE IN COTTON CELL CULTURE The accession numbers for the sequences described in this article are DQ868985 (GhPLA1) DQ868986 (genomic GhPLA1). Supplemental Data The following materials are available in the online version of this article. Supplemental Table S1. The effects of AGPs from cotton calli on the number of explants developing embryogenic calli in cotton cell cultures Supplemental Table S2. The effects of recombinantly expressed GhPLA1 and its phytocyanin-like domain (PL1) on the number of explants developing embryogenic calli in cotton cell cultures ACKNOWLEDGMENTS We acknowledge Dr Sue Finch (Statistical Consulting Centre, Department of Mathematics and Statistics, The University of Melbourne) for assistance with statistical analyses. We also thank Ms Kris Ford (School of Botany, The University of Melbourne) for performing protein sequencing and mass spectrometry analyses.