Grafting Triggers Differential Responses between Scion and Rootstock

Grafting is a well-established practice to facilitate asexual propagation in horticultural and agricultural crops. It has become a method for studying molecular aspects of root-to-shoot and/or shoot-to-root signaling events. The objective of this study was to investigate differences in gene expression between the organs of the scion and rootstock of a homograft (Arabidopsis thaliana). MapMan and Gene Ontology enrichment analysis revealed differentially expressed genes from numerous functional categories related to stress responses in the developing flower buds and leaves of scion and rootstock. Meta-analysis suggested induction of drought-type responses in flower buds and leaves of the scion. The flower buds of scion showed over-representation of the transcription factor genes, such as Homeobox, NAC, MYB, bHLH, B3, C3HC4, PLATZ etc. The scion leaves exhibited higher accumulation of the regulatory genes for flower development, such as SEPALLATA 1–4, Jumonji C and AHL16. Differential transcription of genes related to ethylene, gibberellic acid and other stimuli was observed between scion and rootstock. The study is useful in understanding the molecular basis of grafting and acclimation of scion on rootstock.


Introduction
Grafting is a widely used and traditional method of asexual propagation in fruit crops which do not reproduce true-to-type from seed [1]. The benefits of grafting in vegetable crops are also being recognized in recent years [2]. Rootstocks influence the scion development in several ways, affecting the traits of agricultural interest, such as vegetative vigour, stress tolerance, yield, fruit quality etc. [2,3]. The controlling effect of rootstock over scion is possibly due to altered root-to-shoot and/or shoot-to-root chemical signaling [3]. Several studies on longdistance signaling via graft-union provide evidences for multiple types of mobile signals, such as hormones [4,5,6,7], proteins [8,9], ribonucleoprotein [10], RNAs [11], small RNAs [12,13,14,15,16], minerals [17,18] etc, conferring a wide range of effects on scion development.
Despite the wide use of grafting in agriculture, very little is known about the molecular mechanism of rootstock-regulation of scion's phenotypes. Gene expression studies are useful approaches in understanding the genes involved in the effect of the rootstock. Transcriptional profiling in the scions of Prunus cerasus [19] and Malus domestica [20] revealed differences in

Homografting
Homografting was carried out on young inflorescence stems of A. thaliana plants of uniform age (4-5 weeks) and height (~10 cm), following the procedure described by Nisar et al. [25], with some modifications. The primary inflorescence stem was cut horizontally by using a razor blade, and immediately placed in a petri dish containing sterile water; this part was used as scion for grafting on the same plant. A drop of water was placed on the cut end of the primary inflorescence stem of the same plant, to be used as rootstock. A vertical incision (~1 cm) was made in rootstock and the scion was cut in a wedge shape. Cut ends of the scion and rootstock were attached and wrapped with a parafilm around the graft. A support of a stick was provided to the plant. The plant was covered with a plastic bag to maintain high humidity for three days. The grafting was performed on multiple plants. Three grafted plants showing the best scion growth and development were selected for the study. The newly emerged un-opened flower buds and leaves were harvested from the side branches of scion and rootstock, at the same time. Harvesting of the samples was performed during 10 to 20 days after the graft (DAG). The samples were immediately frozen in liquid nitrogen and stored at -80°C till further use. The experiment was done in three independent biological replicates.

RNA extraction and cDNA synthesis
Total RNA was extracted from the harvested leaf and flower bud samples using Spectrum Plant Total RNA kit (Sigma-Aldrich, USA), following the manufacturer's instructions. On-column DNase (Sigma-Aldrich, USA) treatment was performed as instructed in the manual. The quality and concentration of total RNA were determined by using NanoQuant M200 Pro (Tecan) and agarose gel electrophoresis visualization. Double stranded cDNA synthesis, in vitro transcription to synthesize biotin labeled aRNA, purification and fragmentation of aRNA, and hybridization of arrays was performed following the protocol described in the technical manual of Affymetrix.

Microarray
Affymetrix Arabidopsis ATH1 Genome Array GeneChip was used for microarray experiment. Affymetrix ATH1 GeneChip, a 3' in vitro transcription (3' IVT) expression array, contains more than 22,500 probe sets, representing approximately 24K genes. Labeling and hybridization of ATH1 GeneChips (one sample per chip) was performed according to the manufacturer's instructions (http://www.affymetrix.com/support/technical/manuals.affx). The hybridized arrays were processed by running fluidics script FS450_0004 on an Affymetrix GeneChip Fluidics Station 450 and scanned on Affymetrix GeneChip Scanner 3000. The quality of hybridization was verified according to the Affymetrix microarray standards. The expression console of Affymetrix's GeneChip Command Console (AGCC) software was used for computing cell intensity data of probesets and their positional values from image file. The intensities of probe arrays were normalized by using GeneSpring GX v12 (Agilent Technologies, Santa Clara, USA). The data has been submitted to NCBI (http://www.ncbi.nlm.nih.gov), with accession number GSE61631. Robust Multi-array normalization algorithm (RMA) values of probe sets were used for further statistical analysis. One-way ANOVA analysis was carried out in GeneSpring software with 'Asymptotic' p value computation and Benjamini-Hochberg false discovery rate (FDR) for multiple test correction (at p 0.05). The probe sets satisfying the criteria of p-value ( 0.05) and fold change ( 2) were used as differentially expressed genes for further analysis.
Functional annotation of the differentially expressed probe sets was obtained using the information available at TAIR (http://www.arabidopsis.org/tools/bulk/microarray/index.jsp) and PLEXdb (http://www.plexdb.org/modules/PD_general/tools.php). MapMan software was employed for visualization of differences in gene expression, and enrichment of functional categories in differentially expressed genes using the Wilcoxon rank-sum test (p value 0.05) [27,28].
Enrichment of Gene Ontology (GO) terms in the differentially expressed genes was performed using AgriGO analysis tool (http://bioinfo.cau.edu.cn/agriGO) [29], with Fisher tests and Bonferroni multiple testing correction (p 0.05). Kyoto Encyclopedia of Genes and Genomes (KEGG) categories was assigned by the plant gene set enrichment analysis toolkit (http://structuralbiology.cau.edu.cn/PlantGSEA/analysis.php) with fisher test function.

Quantitative RT-PCR
For the validation of microarray data, quantitative RT-PCR was performed for five randomly selected genes in three biological replicates. cDNA was prepared from 500 ng of total RNA using Transcriptor First Strand cDNA Synthesis Kit (Roche, USA) according to manufacturer's instructions. Gene expression was analyzed using 2X QuantiTect SYBR Green (Qiagen, USA), with a 200 nM primer concentration in a qRT-PCR machine (7500 Fast Real-Time PCR System, Applied Biosystems), according to the manufacturer's instructions. The expression of genes of interest was normalized using housekeeping gene (polyubiquitin 10; At4g05320) and relative change in gene expression was quantified as described previously [30].

Results and Discussion Homografting
Homografting experiments were carried out on young primary inflorescence stems of Arabidopsis plants (Fig 1A). Though difficulties have been encountered in maintaining hydraulic turgor across the graft junction [23], with the recent developments in grafting technique [25], the integrity of the graft union formation has been improved. In a successful flowering stem graft, vascular connection is established by about 7 DAG [25]. The floral stem graft, which maintains shoot apical dominance with a taller primary stem, indicates a functional vascular connection at the graft junction [25]. Three floral stem grafts with the best scion growth and development (Fig 1B), indicating appropriate transport of water, nutrients and signalling molecules across the graft junction, were selected for the study. The longitudinal and transverse sections across the graft-junction (10 DAG) confirmed the establishment of vascular connections between scion and rootstock stems (Fig 1C-1E). Callus proliferation near the wedge junction ( Fig 1D) is indicative of good regenerative growth, confirming the integrity of graft-union [25]. The newly emerged un-opened developing flower buds and leaves were harvested from the side branches of scion and rootstock (10-20 DAG), for gene expression analysis. At maturity, siliques of scion were comparable to that of rootstocks (Fig 1F-1I). However, the slight reduction in silique length and number of seeds could be due to grafting generated effects on scion development.

Homografting alters the expression of many genes in flower buds and leaves
The transcriptional changes were examined in flower buds and leaves of scion and rootstock, emerged during 10 to 20 days after the homograft. A total of 840 genes were identified as differentially expressed, by two folds or more at p 0.05, in flower buds and/or leaves of scion and rootstock (Fig 2, S1 Table). The fold expression of five randomly selected genes was validated by qPCR analysis (S1 Fig). The differentially expressed genes have been further analyzed and discussed.  Table. doi:10.1371/journal.pone.0124438.g002

Divergent profiles of differentially expressed genes in flower buds and leaves
MapMan, AgriGO, and KEGG categorization of the differentially expressed genes revealed a differential level of accumulation of a divergent set of genes in flower buds and leaves of scion as compared to rootstock (Tables 1-6). Up-regulated hormonal metabolism was observed in The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t001 The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t002 flower buds of both scion and rootstock. However, several genes related to hormonal signaling pathways were over-represented in scion buds, conferring tolerance to stress [31]. Higher accumulation of the transcripts for Late Embryogenesis Abundant (LEA) proteins in flower buds of scion could protect cellular proteins from aggregation, under the abiotic stresses such as desiccation, osmotic stresses, temperature, salinity etc [32]. The transcription factors, such as homeobox genes, MADS box and MYB which express in responses to several stresses [33,34,35,36,37], and hormonal stimuli [33,38,39], were up-regulated in flower buds of the scion. In addition, a few genes involved in protein synthesis and photosynthesis were upregulated in the flower buds of scion, as compared to that of rootstock (Table 1). In contrast to scion, genes associated with the MapMan functional categories of cellular processes, cell organization, CHO metabolism, and F-box proteins which are critical for the controlled degradation of cellular proteins, were up-regulated in the flower buds of rootstock (Table 1). This could be indicative of comparatively increased rate of cell division in flower buds of rootstock. Enrichment of Gene Ontology (AgriGO) terms in the differentially expressed genes revealed that most of the genes up-regulated in the flower buds of scion, belonged to the biological process of responses to different stimuli (chemical, abiotic stresses and hormone etc.), localization, transport, and oxidoreductase activities ( Table 3). The GO terms related to microtubule-based movement and motor activities, involved in cytoskeleton organization and developmental processes [33], were enriched in flower buds of rootstock ( Table 3). The leaves of scion showed a comparatively high level of accumulation of genes associated with transport and lipid metabolism (Tables  2 and 4). Rootstock leaves accumulated the transcripts predicted to be involved in responses to stress, biotic and abiotic stimuli, and defense and wound responses (Tables 2 and 4).
The functional significance of genes was also explored by KEGG, which exhibited significant over-representation of genes associated with amino acid metabolism, and biosynthesis of other secondary metabolites such as alkaloids, stilbenoid, diarylheptanoid, and gingerol in flower buds (Table 5), and flavonoid and phenylpropanoid biosynthesis in leaf (Table 6) of scion. Alkaloids have several biological significance of being active stimulators, inhibitors and terminators of growth [40]. The genes could participate in several regulation mechanisms and confer protection against environmental stresses to the plant organs [41]. In rootstock, the genes associated with ascorbate, aldarate and inositol phosphate metabolism were up-regulated in flower buds (Table 5), whereas leaves exhibited active amino acid metabolism, glutathione and alphalinolenic acid metabolism, and plant-pathogen interaction ( Table 6).

Meta-analysis
Similarity search meta-analysis was performed against 3287 diverse collections of Arabidopsis microarray data sets listed in Genevestigator, by using the differentially expressed transcripts between scion and rootstock. The perturbation showing maximum similarity with our data The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t003 comprised the transcriptional comparison between drought study vs control plants in flower buds and leaves (S2 Fig). The meta-analysis suggests that grafting induces drought-type responses in flower buds and leaves of scion, which could play a role in acclimation to grafting induced stresses. The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t004 The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t005 The details of the genes are given in S4 Table. doi:10.1371/journal.pone.0124438.t006

Differential expression of transcription factors in flower buds and leaves
Out of about 25,500 genes, around 2000 transcription factor genes have been recognized in Arabidopsis genome [42]. Transcriptional regulation plays a pivotal role in temporal and spatial control over gene expression in plants. Altered expression levels of transcription factors were observed in flower buds and leaves developed on scion and rootstock (Fig 3).  Table. doi:10.1371/journal.pone.0124438.g003 Homeobox transcription factor family (HB) genes were significantly up-regulated in flower buds (e.g. at2g18550, at2g46680, at4g36740, at5g66700 and at3g61890) and leaves of scion (at2g18550, at2g46680, at5g66700, at1g14440, at1g75240), as compared to rootstock. In several studies, the expression regulation of homeobox genes by different abiotic and biotic stimuli has been inferred [33,34,38,43]. The higher level of transcript accumulation of homeobox genes in the vegetative and reproductive organs of scion could be helpful in achieving tolerance to subsequent stresses after grafting.
MYB transcription factor super family play regulatory roles in differentiation, metabolism and development processes, and defense responses in plants [35]. Scion flower buds showed higher expression of MYB genes which are known to regulate and/or respond during cell cycle (at5g59780) [44], differentiation (trichome initiation) (at3g27920) [46], phenylpropanoide pathway (at1g22640) [46], abiotic stress responses (e.g. drought, light and wounding) (at1g06180) [45]and abiotic and biotic stimuli (at1g71030 and at5g58900) [47]. However, the expression of MYB35/TDF1 (at3g28470), essential for anther tapetum development [48], was observed down regulated by three folds in scion flower buds. This could be indicative of grafting effect on tapetum development. Surprisingly, up-regulation of the MYB genes was not observed in scion leaves at a significant level. On the other hand, rootstock leaves exhibited enhanced expression of the MYB-related genes having role in signal transduction (at5g06800) [49], and express in response to abiotic or biotic stimuli (at5g05790) [47].
The ethylene response factor, AP2-RAP2.6 (at1g43160), was significantly up-regulated in scion flower buds. It functions in plant response to various abiotic and biotic stresses, possibly through ABA-dependent pathway [58,59]. The importance of RAP2.6 has been emphasized in achieving water-stress tolerance in plant tissues [59]. This was further corroborated by visualization of abundance of genes associated with phytohormones, mainly ABA, that could play important roles in mediating responses to various stresses in scion flower buds (Fig 3). Enhanced transcript accumulation of key regulators of sulfur assimilation pathway, LSU2 (at5g24660) [60]and HY5 (at5g11260) [61], could be indicative of implications of stress, presumably water stress [60], on the sulfur assimilation in scion flower buds. Drought conditions affect the regulation of sulfur assimilation in plant tissues, and it has been anticipated as a fertile ground for new discoveries connecting primary sulfur metabolism with the stress responses, mainly drought [62]. The expression elevation of bZIP1 (at5g49450), a positive regulator of plant tolerance to salt, osmotic and drought stresses [63], waterdeficit stress related remorin family protein (at2g41870) [64]and the stress responsive heat shock C1 (at3g24520) [53]anticipates role of these genes in acclimating scion on rootstock.
The E genes of ABCDE model, SEPALLATA (SEP) 1-4, are essential for the normal development of petals, stamens, carpels and sepals [65]. SEP 1-4 genes (at5g15800, at3g02310, at1g24260 and at2g03710) were up-regulated in the leaves of scion, as compared to rootstock. Mutation in SEP genes leads to development of flowers composed of leaf-like organs, whereas, over-expression promotes early flowering without affecting floral morphology [66]. In addition, some other key regulators of flowering, Jumonji C and AHL16, were up-regulated in scion leaves. Jumonji C (at3g20810) is a histone demethylase which regulates the period length in Arabidopsis by chromatin remodeling [67]. AHL16 is an AT-hook DNA binding protein, which regulates vegetative to the reproductive phase transition of the meristem, and flowering time [68]. The enhanced level of transcripts for SEP 1-4, AHL16 and Jumonji C genes could be suggestive of the initiatives taken by scion for regulating flowering. The MYB transcription factor, ATRL2 (at2g21650), was highly expressed in rootstock leaves. The gene has been observed to be involved in stress responses [53,69], besides in ovule development and control of floral asymmetry [70]. Rootstock leaves showed elevated transcription of genes related to ethylene (at4g17500, at5g47220, at4g17490, at4g32800, at5g07580 and at1g13260) and gibberellic acid (at1g66350 and at5g56300). Thus, enhanced expression was observed for the genes related to ethylene and gibberellic acid in scion flower buds and rootstock leaves (Fig 4). Highly expressed genes A total of 18 genes (other than transcription factors) were identified with an expression difference (scion vs rootstock) of at least 8 folds in flower buds and leaves ( Table 7). The transcript level of dehydration and ABA inducible genes-ABA responsive protein-related (at3g02480) [56], Histone H1-3 (at2g18050) [71], Responsive to ABA 18 dehydrin family protein (at5g66400) [71]and protein phosphatase 2C (at1g07430 and at3g11410) [72,73], were higher by many folds in scion flower buds and leaves (Table 7). This indicates existence of ABAmediated signaling in scion which could help the scion organs in withstanding grafting related stress, such as water deficiency. This is further supported by higher level of expression of Late Embryogenesis Abundant-7 protein in scion buds which confers resistance to abiotic stresses and ABA sensitivity [74]. The grafting may induce ethylene production in scion flower buds, as suggested by the several fold up-regulation of 1-Aminocyclopropane-1-carboxylate (ACC) oxidase which is involved in the final step of ethylene production in plant tissues [75,76]. Expression of Cysteine Endopeptidase 1 was also found highly expressed, which could be related to ethylene regulation [77]or in response to stress stimuli [78]in scion flower buds. Another ethylene induced and flowering related gene, Xyloglucan Endotransglucosylase/hydrolase3 [79], was up-regulated in scion flower buds. In scion buds, no gene was down-regulated with a difference of at least 8 folds.

Conclusion
Grafting triggers differential expression of numerous genes related to stress, biotic and abiotic stimuli, hormonal pathway, and flowering etc. in flower buds and leaves of the scion and rootstock. The study is useful in understanding the molecular basis of grafting and the intermediates involved in the acclimation of scion on rootstock.
Supporting Information