Stronger sink demand for metabolites supports dominance of the apical bud in etiolated growth

Abstract

The potato tuber is a swollen underground stem that can sprout under dark conditions. Sprouting initiates in the tuber apical bud (AP), while lateral buds (LTs) are repressed by apical dominance (AD). Under conditions of lost AD, removal of tuber LTs showed that they partially inhibit AP growth only at the AD stage. Detached buds were inhibited by exogenous application of naphthaleneacetic acid (NAA), whereas 6-benzyladenine (6-BA) and gibberellic acid (GA ₃ ) induced bud burst and elongation, respectively. NAA, applied after 6-BA or GA ₃ , nullified the latters’ growth-stimulating effect in both the AP and LTs. GA ₃ applied to the fifth-position LT was transported mainly to the tuber’s AP. GA ₃ treatment also resulted in increased indole-3-acetic acid (IAA) concentration and cis-zeatin O-glucoside in the AP. In a tuber tissue strip that included two or three buds connected by the peripheral vascular system, treatment of a LT with GA ₃ affected only the AP side of the strip, suggesting that the AP is the strongest sink for GA ₃ , which induces its etiolated elongation. Dipping etiolated sprouts in labeled GA ₃ showed specific accumulation of the signal in the AP. Transcriptome analysis of GA ₃ ’s effect showed that genes related to the cell cycle, cell proliferation, and hormone transport are up-regulated in the AP as compared to the LT. Sink demand for metabolites is suggested to support AD in etiolated stem growth by inducing differential gene expression in the AP.

Apical dominance, auxin, cytokinin, etiolated growth, gibberellin, potato

Introduction

The potato ( Solanum tuberosum L.) tuber is a swollen underground stem that can autonomously sprout under dark conditions when dormancy is released, exhibiting apical dominance (AD). Post-harvest storage of tubers in a dark and cold environment delays sprouting, and tuber AD is gradually lost with physiological aging (reviewed by Eshel and Teper-Bamnolker, 2012 ). Lateral buds (LTs) are inhibited by the apical bud (AP) only after its autonomous dormancy is released ( Teper-Bamnolker et al. , 2012 ).

Hormonal control of tuber dormancy release and sprouting in the dark has been investigated in several studies, but is still not well understood (see review by Sonnewald and Sonnewald, 2014 ). Mainly abscisic acid (ABA) and ethylene have been linked to the onset and maintenance of tuber dormancy ( Suttle and Hultstrand, 1994 ; Suttle, 2004 b ): ABA levels are highest in deeply dormant tubers and decline during storage, as dormancy is released ( Biemelt et al. , 2000 ; Destefano-Beltran et al. , 2006 ); ethylene has been recognized to play a dual role in sprouting, promoting or inhibiting etiolated bud growth at low and high concentrations, respectively ( Suttle, 1998 , 2004 b , 2009 ; Pierik et al. , 2006 ). Gibberellins (GAs) and cytokinins (CKs) have been associated with dormancy release and etiolated sprouting of post-harvest potato tubers or their detached buds ( Suttle, 2004 a ; Hartmann et al. , 2011 ; Rentzsch et al. , 2011 ). By applying the synthetic CK 6-benzylaminopurine (BAP), Hartmann et al. (2011) showed that CK stimulates bud burst while GA ₃ application is needed to induce further bud growth. Strigolactones (SLs) have been shown to nullify the sprouting-inducing activity of CK and GA ( Pasare et al. , 2013 ). Exogenous application of auxin, either as indole-3-acetic acid (IAA) or 1-naphthaleneacetic acid (NAA), has dose-dependent activity and can either inhibit or stimulate tuber bud growth when applied at high or low doses, respectively ( Hemberg, 1985 ; Suttle, 2007 ).

Metabolic competence has been suggested as a prerequisite for sprouting initiation. A rapid shift in the tuber during post-harvest storage from storage metabolism (starch synthesis) to consumption metabolism has been suggested to resemble the transition from sink to source, consistent with bud growth ( Sonnewald, 2001 ). Viola et al. (2007) suggested that tuber bud growth is initially prevented by substrate limitation mediated via the symplastic connection of each bud. Recently, sugars, rather than auxin, have been shown to be necessary and sufficient for regulating the very early stages of bud outgrowth in pea plants following decapitation ( Mason et al. , 2014 ). The demand for sugars by the intact shoot tip prevents the initial outgrowth of axillary buds, and overrides the effect of auxin.

It is still unclear whether hormone relationships and/or sugars produced in the parenchyma regulate sprouting timing and AD of potato tubers in the dark. The potato tuber loses AD with physiological aging or as a result of exposure to abiotic stress ( Eshel and Teper-Bamnolker, 2012 ). In this study, the existence of bidirectional signaling between buds is demonstrated, suggesting a sink–source relationship between the AP and LTs that mediates the tuber’s etiolated sprouting shape.

Materials and methods

Plant tissue

Potato ( Solanum tuberosum L.) mini-tubers, cv. Nicola, were used in all experiments. Explants were propagated on Murashige and Skoog (1962) medium with 3% (w/v) sucrose (Duchefa, Haarlem, Netherlands), MS vitamins (Sigma-Aldrich, Rehovot, Israel), and 8g l ⁻¹ agar (Difco, St. Louis, MO) in magenta boxes (Sigma-Aldrich) containing 30–35ml medium. Cultures were incubated at 25 °C with 75 μmol m ⁻² s ⁻¹ cool white fluorescent light. Explants were transferred to a controlled-atmosphere greenhouse at 25 °C, and mini-tubers were harvested after 90–100 d and stored at 4 °C and 95% relative humidity. All 14 °C incubations were performed at 83% relative humidity.

Detached-bud assay

The bioassay was performed according to a previously described procedure ( Hartmann et al. , 2011 ) with some modifications: tubers were washed in tap water and dried at room temperature, then the AP or LTs were excised using a 0.5-cm diameter cork borer and washed three times for 10min in sterile filtered buffer at pH 6.5 containing 20mM MES, 300mM D-mannitol, and 5mM ascorbic acid. Tissue cylinders were incubated with sterile water (control), 100 µM gibberellic acid (GA ₃ ; Sigma), 50 µM 6-benzyladenine (6-BA; Sigma), or 50 µM NAA (Sigma) for 5min and placed in a 48-well plate (Nunc, Roskilde, Denmark) lined with moist filter paper (Whatman, Hillsboro, OH). In the combined treatments, buds were treated with GA ₃ or 6-BA as described above or 1mM 1-N-naphthylphthalamic acid (NPA) and, in some of the treatments, 1h later with NAA for 5min. Plates were stored at 14 °C in the dark. The filter paper was moistened every 48h by adding sterile water.

Tissue-strip assay

Tuber tissue strips were sliced at a depth of 1cm using a scalpel. Dual-bud strips contained an AP on one side and a LT on the other; triple-bud strips contained one AP either centered between two LTs or at the edge of the strip. Strips were handled as described for the detached-bud assay and treated by dipping only the treated bud in solution for 10min, transferring to a Petri dish lined with moist filter paper, and storing in the dark at 14 °C.

Hormone analysis

Tubers were washed in tap water and dried at room temperature. Treatment with exogenous phytohormones was performed according to Suttle (2009) with some modifications: the treated bud base was stabbed to a depth of 3mm with a 1-mm diameter sterile tip, 2mm from the fifth LT (counted from the AP down the tuber) and then subjected to vacuum infiltration for 10min. The AP and sixth LT (middle of the tuber and opposite to the treated fifth LT) were sampled at 0, 24, 48 and 72h after treatment using a 0.3-mm diameter cork borer into N ₂ and stored at –80 °C until use.

Frozen tissue was dried and ground with a mortar and pestle, using 50mg per treatment in three replicates. Phytohormones (ABA and ABA metabolites, CKs, auxins, and GAs) were quantified at the National Research Council of Canada’s Plant Biotechnology Institute (Saskatoon, SK) by ultra-performance liquid chromatography–electrospray ionization mass spectrometry (UPLC–ESI-MS/MS) ( http://archive.nrc-cnrc.gc.ca/eng/facilities/pbi/plant-hormone.html ) following a previously described procedure ( Chiwocha et al. , 2003 , 2005 ). MassLynx™ and QuanLynx™ (Micromass, Manchester, UK) were used for data acquisition and analysis, respectively. Statistical analysis of the data was performed with JMP-in software (version 3 for Windows; SAS Institute, Cary, NC, USA).

GA ₃ –fluorescein (FL) accumulation in tuber sprouts

Fluorescently labeled GA ₃ (GA-FL; Shani et al. , 2013 ) was used to detect the specific site of GA ₃ accumulation in the tuber sprout meristems and the AP sink. Labeling tuber strips with GA-FL caused high background noise, probably due to the high starch content. Apical sprouts, 6cm long with two potential branching nodes, were detached from the tuber and immediately immersed in a 1.5-ml solution of 20 µM GA–FL or FL. Stems were vacuum-infiltrated in 50-ml tubes for 1min, then incubated for 30h in the dark at 14 °C. The AP and LT were sliced from the stem and fixed for 48h in formalin/acetic acid/alcohol (FAA) solution in the dark at 4 °C, then rinsed five times in 1% (v/v) phosphate buffered saline (PBS) and vacuum-infiltrated overnight with a 25% sucrose solution. The tissue was rinsed in 1% PBS, dried and frozen overnight at –20 °C. The frozen bud was sliced into 50-µm thick sections with a Cryostat CM3000 (Leica Microsystems, Wetzlar, Germany). Bud sections were imaged in a laser-scanning confocal microscope (Leica SP8, Leica Microsystems), with the laser set at 488nm for excitation and 500–540nm for FL-derivative emission. Image analysis and signal quantification were performed with the ROI (region of interest) analysis function of Fiji software ( Schindelin et al. , 2012 ). Average signal intensity from a 5-circle ROI (5×5mm) was quantified from five sections of five different sprouts.

RNA-Seq and bioinformatics

The transcriptome of the AP and LT of GA ₃ -treated tubers was sequenced using Illumina HiSeq 2000 and Trueseq protocols, at the Genome Center, Life Sciences and Engineering, Technion, Israel. Eighteen libraries were generated representing tissue from the LT in the sixth position and the AP at 0, 24 and 72h after GA ₃ injection into the LT located in the fifth position. These libraries were subjected to single-end RNA-Seq, generating up to 100-nucleotide-long sequences. The transcriptome datasets are available at the NCBI Sequence Read Archive (SRA) under accession number SRP059022 and BioProject accession PRJNA285785.

After cleaning the data with Trimmomatic software version 0.32 ( Bolger et al. , 2014 ) to remove low-quality reads, sequences were aligned to the S. tuberosum genome version PGSC_DM_v3.4 ( PGS Consortium, 2011 ), downloaded from the Sol Genomics database ( http://solgenomics.net ; Fernandez-Pozo et al. , 2015 ), using Bowtie2 alignment software ( Langmead and Salzberg, 2012 ). The abundance estimates of the number of reads per gene were performed using RSEM software ( Li and Dewey, 2011 ). Differential expression analysis was performed with the edgeR package ( Robinson et al. , 2010 ), using a threshold false discovery rate (FDR) <0.001 ( Benjamini and Hochberg, 1995 ) and fold-change greater than 4 or lower than 1/4. R Bioconductor ( Gentleman et al. , 2004 ) was used to calculate hierarchical clustering of gene expression and to visualize heat maps. The normalized expression value was centered and log ₂ -transformed for visualization purposes with a script taken from the Trinity pipeline ( Haas et al. , 2013 ). Differential expression was plotted as heat maps using the pheatmap package ( https://cran.r-project.org/web/packages/pheatmap/index.html ).

Blast2GO ( Conesa et al. , 2005 ) was used for annotation. Gene ontology (GO) enrichment analysis was performed using Fisher’s exact test ( Upton, 1992 ) with multiple testing correction of FDR ( Benjamini and Hochberg, 1995 ). The MapMan3.6.0RC1 tool ( Thimm et al. , 2004 ) was used with mapping file S. tuberosum (PGSC_DM_v3.4) to display expression profiles at the pathway level. ‘Venny’ ( http://bioinfogp.cnb.csic.es/tools/venny/ ) was used for Venn diagram construction.

Results

Apical–Lateral bud cross-talk in potato tuber

The inhibitory effect of the dominant AP on LTs development is well-defined in plants ( Sachs and Thimann, 1964 ). Removal of the dominant AP usually induces branching of most stems. However, the contribution of the LTs to AP development, to the best of our knowledge, has never been described. Experiments were performed to examine the existence of a bidirectional interaction between the potato AP and LTs. Dormant non-sprouting tubers were stored at 4 °C until they reached the four different physiological stages with respect to dormancy and sprouting condition described in Fig. 1 : Dormant – dormant tuber (more than 2 weeks are needed for the AP to sprout at 14 °C); Active – tuber with dormant LTs (less than 2 weeks are needed for only the AP to sprout at 14 °C); AD – sprouting tubers with AD where LTs are suppressed by the AP; AD loss – sprouting tuber that has lost its AD, with several developing LTs. Tubers at all four stages were transferred to 14 °C and divided into three treatments: non-treated (Control), wounding next to all LTs (Wound), and removal of all LTs (LTs removed). Following the treatments, the length of the AP was measured for up to 50 d of incubation.

Fig. 1.

Effect of lateral bud (LT) removal on apical bud (AP) development. ‘Nicola’ potato tubers at four different physiological stages were compared at 14 ºC: Dormant , dormant tuber (more than 2 weeks needed for the AP to sprout at 14 ºC); Active , tuber with dormant LTs (less than 2 weeks are needed for only the AP to sprout at 14 ºC); AD , sprouting tubers with apical dominance (AD) in which LTs are suppressed by the AP; AD loss , sprouting tuber that has lost its AD, with several developing LTs. Control, non-treated tubers; Wound, wounding 0.5cm from each lateral bud; LT removed, removal of all lateral buds. Bars indicate ±SE of three repeats, n (tuber number in each repeat) = 25.

Surprisingly, the APs of AD tubers with LTs removed developed faster than those of wounded tubers or non-treated controls, reaching an average of 2.5cm after 39 d of incubation as compared to 1.85 and 1.75cm, respectively ( Fig. 1 ). This effect was unique to AD tubers. None of the other groups of tubers – Dormant , Active , or AD loss – showed any significant effect of LTs removal on AP growth ( Fig. 1 ). This suggested that the non-dormant LTs have an inhibitory effect on AP growth only at the AD stage.

Effect of hormones on detached-bud and whole-tuber sprouting

Based on studies on phytohormone transport in other plants (reviewed by Depuydt and Hardtke, 2011 ), the bidirectional effect of the AP and LTs in potato tubers has been hypothesized to be a result of hormone transport. To reveal the possible role of plant hormones in the cross-talk between LTs and the AP, their differential effect on isolated buds was examined. Detached AP or LTs from AD tubers were treated with NAA, 6-BA, or GA ₃ . There was no visible effect of NAA, 7 d after exogenous hormone application, whereas 6-BA and GA ₃ induced bud burst and elongation of the primary meristem covered by the scale leaves ( Fig. 2 ). Combined application of either 6-BA or GA ₃ with NAA led to inhibition of bud burst and elongation, resulting in a phenotype similar to that observed following NAA treatment ( Fig. 2 ). This suggested that NAA, although applied 1h after 6-BA or GA ₃ , could eliminate their growth-regulating effect. Application of the same hormonal treatments to detached LTs had a similar effect, suggesting that AP growth priority is not related to its sensitivity to phytohormones ( Fig. 2 ).

Fig. 2.

Effect of exogenous phytohormone treatment on development of apical (AP) and lateral (LT) buds detached from ‘Nicola’ potato tubers. Control, application of DDW with 0.0001% v/v DMSO and 0.001% w/v NaOH; NAA, 50 µM α-naphthalenacetic acid; 6-BA, 50 µM 6-benzylaminopurine; GA ₃ , 100 µM gibberellic acid (scale bars = 1000 μm); 6-BA+NAA or GA ₃ +NAA, NAA was applied 1h after 6-BA or GA ₃ (scale bars = 500 μm). Images represent the average response of 48 buds in each treatment, 7 d after application.

To determine whether non-detached buds respond similarly to the phytohormones, three phytohormones were applied to the tuber, to either the AP or the fifth LT (located in the fifth position from the AP down the tuber), and bud response was examined. The response of the non-detached AP to NAA, 6-BA, and GA ₃ was similar to that observed for the detached AP ( Fig. 3 ). Interestingly, NAA, 6-BA, and GA ₃ applied to the non-detached LT affected only the AP of the tuber, inducing growth inhibition, bud burst only, and bud burst and elongation, respectively ( Fig. 3 ). The most surprising effect was of application of GA ₃ to the fifth LT, which induced early AP burst followed by elongation, suggesting that GA ₃ moves in the tuber vascular system and/or affects other endogenous phytohormone levels.

Fig. 3.

Effect of exogenous phytohormones, applied to the apical bud (AP) vs. lateral bud (LT) of ‘Nicola’ potato tubers. Left and right columns represent treatment of the AP or fifth position LT, respectively. Control, application of DDW with 0.0001% v/v DMSO and 0.0001% w/v NaOH; NAA, 50 µM α-naphthalenacetic acid; 6-BA, 50 µM 6-benzylaminopurine; GA ₃ , 100 µM gibberellic acid. Black arrowheads indicate treatment location (scale bar = 2cm). Images represent the average response of 90 tubers per treatment, 14 d after application.

Effect of exogenous GA ₃ treatment on endogenous phytohormone levels

To determine GA ₃ ’s effect on hormonal fluctuations within tuber buds, it was applied to the tuber’s fifth bud (LT5; Fig. 4A ) and levels of GA, ABA, auxin, and CK in both the AP and the sixth LT (located opposite to the fifth bud on the tuber; LT6; Fig. 4A ) were analyzed by UPLC–ESI–MS/MS (see Supplementary Table S1 at JXB online). Since natural GA ₃ cannot be detected in potato tuber tissue ( Suttle, 2004 a ; Hartmann et al. , 2011 ), any measured GA ₃ should originate from the applied GA ₃ . Within 24h after treatment of the fifth LT, the GA ₃ level increased from non-detectable to 28ng g ⁻¹ DW in the AP and remained at around this level until 72h after application ( Fig. 4B ). Increasing levels of GA ₃ were detected in the sixth LT as well, although they were lower than those measured in the AP, reaching 9 and 14ng g ⁻¹ DW after 24 and 72h, respectively ( Fig. 4B ). This finding indicated transport of GA ₃ from the fifth LT mainly toward the AP, and at lower strength toward the sixth and possibly other LTs.

Fig. 4.

Phytohormone levels in the apical bud (AP) vs. sixth lateral bud (LT6) of ‘Nicola’ potato tubers 0, 24, 48 and 72h after treatment with 100 µM gibberellic acid (GA ₃ ) applied to the fifth LT (LT5). (A) Schematic presentation of cut length of the treated tuber. The triangles, dashed lines and the outer solid line represent the buds, vascular ring, and the tuber skin, respectively. (B–D) Contents of (B) GA ₃ , (C) indole acetic acid (IAA), and (D) cis-zeatin O-glucoside (Cis-ZOG). Data represent averages ±SE of three repeats, n =90. Different uppercase letters denote significant differences for each time point in the same bud (AP or LT); different lowercase letters denote significant differences for each time point in the AP as compared to LT ( P <0.05).

Free IAA was the only auxin detected in the tuber buds, before and after GA ₃ treatment, whereas no IAA-amide or indole-3-butyric acid could be detected (see Supplementary Table S1 ). Initial free IAA levels were higher in the sixth LT than in the AP ( Fig. 4C ); however, following the treatment, this relationship reversed. IAA concentration did not change significantly in the AP up to 72h after GA ₃ treatment and stabilized at about 140ng g ⁻¹ DW. The initially high IAA levels in the LT decreased in the first 48h to 92ng g ⁻¹ DW and then rose to 138ng g ⁻¹ DW 72h after treatment, equalizing with the AP levels ( Fig. 4C ). This suggested transport of IAA between the AP and LTs in either direction, or local IAA biosynthesis induced by GA ₃ .

Treatment of the LT with GA ₃ induced an increase in the CK derivative cis-zeatin O-glucoside (Cis-ZOG) in the AP, from 36ng g ⁻¹ DW 0 and 24h after treatment, to about 82ng g ⁻¹ DW 72h after application ( Fig. 4D ). At the same time, a minor change in Cis-ZOG was observed in the LT. The specific increase in the AP suggested a possible role of Cis-ZOG in the developmental priority dominance of the AP in response to GA ₃ treatment. Other CK derivatives – (trans) zeatin-O-glucoside (t-ZOG), (trans) zeatin (t-Z), (cis) zeatin (c-Z), dihydrozeatin, (trans) zeatin riboside, (cis) zeatin riboside, dihydrozeatin riboside, isopentenyladenine, and isopentenyladenosine – were below the limit of detection (see Supplementary Table S1 ).

Transport of phytohormones between buds of potato tubers

Slicing a potato tuber vertically, bud to stem, shows the xylem vascular system, starting from the stolon and connecting all buds to the inner medulla ( Fig. 5A ). Buds are connected by a vascular ring as well, located under the potato skin, which probably functions as phloem ( Fig. 5A ). To determine phytohormone transport between potato tuber buds, a dual-bud tissue-strip assay was developed. This assay enabled the study of the specific transport between any given tuber-bud combination connected by only the vascular ring, the upper cortex, and the tuber skin ( Fig. 5A ). Tissue strips that contain the AP on one side and the fifth LT on the other, with no additional buds, were detached and treated with exogenous phytohormones. Treatment on the AP side of the strip with 6-BA induced AP burst, whereas GA ₃ application resulted in AP burst and elongation. Both hormones had the same effect on the AP in the strip system as in isolated detached single-buds. Interestingly, when dual hormonal treatment was performed in the strip system – treatment of the AP with either 6-BA or GA ₃ and of the LT with NAA – AP growth was completely inhibited, suggesting that NAA can be transported in the vascular ring from the LT to the AP, and eliminate the local effect of 6-BA or GA ₃ ( Fig. 5B ). Treatment on the LT side of the strip with 6-BA or GA ₃ induced bud burst in both buds (LT and AP) but elongation, induced by GA ₃ was more pronounced on the AP side of the strip ( Fig. 5B ), suggesting movement of the hormone or its signal to the AP sink. Treatment of the AP with NAA eliminated this effect of 6-BA or GA ₃ on the AP ( Fig. 5B ).

Fig. 5.

Phytohormone transport after exogenous application in dual-bud tissue strips sliced from ‘Nicola’ potato tubers. (A) Schematic representation of the tissue-slicing method; stolon, represents the stolon connecting point; LT and AP triangles, represent lateral and apical buds; straight lines indicate the inner vascular system starting from the stolon and connecting all buds; dashed lines and the outer solid line represent the vascular ring and tuber skin, respectively. (B) Tissue strips 10 d after application; the phytohormone applied is marked on the side that was treated. Images represent the average response of 45 tissue strips per treatment.

The apical bud phytohormone sink and effect on auxin transport

The higher sensitivity of the AP to phytohormone treatment observed in the strip system could be due to its being a stronger sink. To examine this possibility, a triple-bud strip system was utilized. Triple strips were prepared with two types of bud combinations: the AP between the two LTs, and the AP at the edge of the strip ( Fig. 6A ). By adding GA ₃ to one of the LTs, located at the strip edge, and forcing it to pass through the central bud (AP or LT), the sink strength of AP vs. LTs can be determined. The AP was found to be most affected by the GA ₃ treatment independent of its position on the strip, sprouting earlier and more vigorously than the other buds. This suggested that it is the strongest sink, whether it is in the center or at the edge of the strip ( Fig. 6A , B ).

Fig. 6.

Comparing phytohormone sink strength of the apical bud (AP) to that of lateral buds (LTs). Detached tri-bud tissue strips sliced from ‘Nicola’ potato tubers, 10 d after phytohormone treatments (100 µM gibberellic acid, GA ₃ ). (A) The treatment is indicated on the side to which it was applied (Con, control). Black arrows indicate the position of the AP; the scale bar is 2cm. (B) Length of AP and LTs in the non-treated (control) and GA ₃ -treated strips. ‘Center’ or ‘Edge’ indicates the position of the AP on the strip. Data are averages ±SE of three repeats of 15 strips, n =45.

To further localize the exogenous GA ₃ in the tissue, FL-labeled GA ₃ molecules, which have been shown to be bioactive in Arabidopsis and tomato–GA growth assays ( Shani et al. , 2013 ), were used. GA–FLs are stable in plants and promote association between the GA-perception factors GID1 and RGA in vitro, suggesting that GA ₃ –FL surrogates can be utilized to study dynamic GA transport in plants ( Shani et al. , 2013 ).

Detached etiolated stems that contained an AP and two nodes (potential LTs) were cut and their base was dipped in GA ₃ –FL solution for 30h in the dark at 14 °C. AP and LTs sampled from the stem were fixed and observed by confocal microscopy. The GA ₃ –FL signal accumulated specifically in the AP (in the meristem and primordial cells), whereas the control treatment using only FL accumulated in both the AP and LTs simultaneously ( Fig. 7A ). Comparison of fluorescence intensity between the AP and LTs showed no difference with the FL treatment, whereas in the GA ₃ –FL-treated sprouts, the signal was higher in the AP than in the LTs of the same sprout ( Fig. 7B ). Most of the fluorescent GA ₃ –FL signal could be detected in the leaf primordia surrounding the apical meristem. These results, although obtained from a detached sprout, strengthened our hypothesis that the AP is the strongest sink for GA ₃ .

Fig. 7.

Accumulation of gibberellic acid (GA ₃ )–fluorescein (FL) in the apical bud (AP) vs. lateral buds (LTs) of etiolated stem of potato cv. Nicola. (A) Confocal images of fluorescence distribution in the bud: top, 20 µM FL; bottom, 20 µM GA ₃ –FL. Left set, AP; right set, LT. Each meristem (m) is presented in bright field and fluorescence; p, primordial leaf. Scale bar = 500 μm. (B) Relative fluorescence intensity in FL- and GA ₃ –FL-treated AP and LT. Quantification of FL or GA ₃ –FL fluorescence intensity in treated buds under the central zone in each meristem. Data are averages ±SE (five sprout images, three sampling points; n =15). Fluorescence intensity was normalized relative to a background point on the meristem. RFU, relative fluorescence units. (This figure is available in color at JXB online.)

Transcriptomics analysis of the GA ₃ effect on apical vs. lateral buds

To assess differential gene expression in the AP vs. LT, which might be associated with making the AP a stronger GA ₃ sink, the GA ₃ effect on the buds’ whole transcriptomes was analyzed. RNA was sampled from the AP and the sixth LT at three time points (0, 24 and 72h) after GA ₃ injection into the fifth LT. A Venn diagram of AP vs. LT genes’ expression showed that lower expressed genes were mostly found 72h after GA ₃ treatment, whereas higher expressed genes were decreasing over time ( Fig. S1 ).

Overall, 5715 differentially expressed genes (DEGs) were identified. Fragments per kilobase of transcript per million fragments mapped (FPKM values) were centered and then subjected to log ₂ -transformation for heat-map visualization, according to the Trinity software protocol ( Haas et al. , 2013 ). Analysis of the heat map showed four clusters of co-expressed genes ( Fig. 8A , labeled in cyan, red, green and blue, respectively). The four clusters showed different GO-enrichment profiles, revealing the differential effect of GA on the AP vs. LT (for full clusters’ GO classification list, see Supplementary Table S2 ). The profile of the cyan cluster (with 1768 genes) was characterized by a substantial rise in gene expression, but the initial values were positive in the AP and negative in the LT ( Fig. 8B ). Most of included genes were related to the cell cycle, cell proliferation, and hormone transport (see Fig. S2A for GO enrichment terms). The red cluster (with 2511 genes) – the biggest cluster of co-expressed genes – was characterized by a minor decrease in expression level in the AP and constant expression in the LT ( Fig. 8B ). Genes in this cluster were related to sugar and carbon metabolism, photosynthesis, energy, and responses to hormone ( Fig. S2B ). The green cluster (with 417 genes) was characterized by genes that are highly expressed in the LT and have low expression in the AP ( Fig. 8B ); these genes were mainly associated with protein degradation and proteolysis, along with genes regulating these processes ( Fig. S2C ). Finally, the blue cluster (with 1019 genes) was characterized by genes with substantially decreased expression in the AP where their expression was initially high, and only a moderate decrease in expression in the LT where it was initially low ( Fig. 8B ). Most of genes in this cluster were involved in stress responses, as well as in defense and programmed cell death ( Fig. S2D ).

Fig. 8.

Cluster heat map and co-expressed genes after GA ₃ treatment applied to the lateral bud (LT) in the fifth position counting down from the apical bud (AP). Measurements were performed at 0, 24 and 72h at the AP and sixth LT. (A) Heat map of the expression profile of 5715 differentially expressed genes; their normalized expression value was centered and log ₂ -transformed for visualization purposes with a script taken from the Trinity pipeline ( Haas et al. , 2013 ). (B) Clustering was performed using hierarchical clustering with the hclust function in R ( www.r-project.org ). The four main clusters are represented in cyan, red, green, and blue. FPKM, fragments per kilobase of transcript per million mapped reads. (This figure is available in color at JXB online.)

Overview of gene expression related to hormonal regulation

To assess the changes in genes assigned to phytohormone categories, the four clusters in MapMan were analyzed under the regulation-overview pathway. The analysis referred to genes whose expression was associated with five major hormones (ABAs, CKs, GAs, ethylene, and auxins; Fig. 9 ). Genes related to the response to ABA had higher expression patterns in the LT, especially before treatment, and some of the genes, such as ABF , PBR , and HVA22 , were down-regulated in response to GA ₃ treatment. ABA-metabolism-related genes showed generally low expression in the AP and higher expression in the LT, as reflected by the expression of 9-cis-epoxycarotenoid dioxygenase ( NCED ) 3 , 4 , and 6 family genes ( Fig. 9 ).

Fig. 9.

Heat map of gene-expression profiles related to hormonal regulation, after GA ₃ treatment applied to the lateral bud (LT) in the fifth position counting down from the apical bud (AP). Measurements were performed at 0, 24 and 72h at the AP and sixth LT. The MapMan3.6.0RC1 tool with mapping file (PGSC_DM_v3.4) was used to display expression profiles at the pathway level. The figure was generated with pheatmap: pretty heatmaps: R package version 0.7.7 ( https://cran.r-project.org/web/packages/pheatmap/index.html ).

As previously reported, CK is essential for bud activation in potato tubers ( Hartmann et al. , 2011 ). The gene homologs of the CK receptor, Arabidopsis histidine kinase ( AHK ), showed higher expression in the AP, supporting higher sensitivity of the AP to CK signaling. Accordingly, CK-metabolism genes such as cytokinin oxidase ( CKX ) 1 and 7 and isopentenyltransferase ( IPT ) 1 , 3, 4 , and 5 were more highly expressed in the LT than the AP ( Fig. 9 ).

In the GA-response pathway, generally higher expression of GA-metabolism-related genes, such as GA oxidases ( GA20ox and GA2ox ) and GA-responsive genes such as GAST1 , GIP , and RSI (see Supplementary Table S2 ), was detected mainly in the AP, supporting the possibility of induction by GA ₃ mainly via transport to the AP. GA20ox and GA2ox were up-regulated and down-regulated in the AP, respectively, despite the GA ₃ treatment ( Fig. 9 ). This is not in agreement with the observed effect of these genes’ manipulation in transgenic potato lines, where overexpression of the Arabidopsis GA20ox resulted in increased growth, but the converse did not apply, whereas overexpression of GA2ox produced a dwarf phenotype, as is typical for increased and reduced GA content, respectively ( Carrera et al. , 2000 ; Hartmann et al. , 2011 ). We suspect that in the present case increase in GA20ox expression may be related to other factors associated with high sink strength rather than the GA ₃ treatment. Initially high expression of the GA receptor GID in both AP and LT tissue and its down-regulation following GA ₃ treatment were detected ( Fig. 9 ). This is in agreement with previous studies suggesting that the GA signal is perceived by GID1, enhancing its interaction with DELLA, a transcription factor that represses GA signaling ( Sun, 2010 ).

Analysis of ethylene-responsive and metabolism-associated genes revealed higher expression of ethylene receptor-encoding genes 1 , 2 , and 10 , especially in the LT, 72h after GA ₃ treatment. 1-Aminocyclopropane-1-carboxylate synthase ( ACS ), and 1-aminocyclopropane-1-carboxylate oxidase ( ACCO ) genes (see Supplementary Table S2 ), both associated with ethylene biosynthesis, generally exhibited high expression in the AP before treatment (time 0); this expression was down-regulated following GA ₃ treatment. The ethylene-responsive genes showed initially (time 0) high expression in the AP, especially genes whose products are an ethylene-responsive factor such as wound-responsive AP2-like factor ( WRAF ) 2 and 3 ( Sasaki et al. , 2007 ), but later showed reduction to levels similar to those in the LT tissue ( Fig. 9 ).

Analysis of auxin-related genes showed that plant-specific pin-formed ( PIN ) 2 and 3 (see Supplementary Table S2 ) are up-regulated in the AP 72h after GA ₃ treatment. GA ₃ ’s effect on PIN 5 and 6 expression in the LT was more moderate. Higher expression of auxin-responsive genes was detected mainly in the AP, most of them before the GA ₃ treatment. These included the IAA/AUX , small auxin up RNA ( SAUR ) and GH3 auxin-responsive gene families (see Supplementary Table S2 ). Down-regulation of these genes was observed 24h after GA ₃ treatment in most cases, and some of the SAUR s were up-regulated again 72h after treatment. Most of the auxin-responsive genes had low to undetectable levels of expression in the LT ( Fig. 9 ).

Discussion

Lateral bud effect on apical development

In storage, the potato tuber acts like a classical swollen stem; when dormancy is released, AD is determined by the tuber’s age (reviewed by Eshel and Teper-Bamnolker, 2012 ). This suggests that tuber buds stored in the dark are not autonomous, but may interact amongst themselves or with the parenchyma tissue before and after sprouting. Indeed, a previous study showed that AP removal can induce early sprouting of the LTs, suggesting classical stem AD behavior in potato tuber ( Teper-Bamnolker et al. , 2012 ). In this study, removing the LTs of a non-dormant tuber at an early sprouting stage, when AD is still in place, induced faster growth of the AP, suggesting bidirectional interactions between the LTs and AP ( Fig. 1 ). Since removal of dormant LTs did not affect AP growth, the effect may be mediated by competition for growth energy, which is reduced upon LT removal. However, the outcome of the wounding control treatment eliminated this as a cause for enhanced growth of the AP ( Fig. 1 ). A novel concept is suggested in which, during tuber AD, AP and LT growth is controlled by hormones and/or other substance that move between these sites.

Hormonal homeostasis in response to GA ₃ treatment

The function of the plant hormone auxin is central to most established AD models for plants grown in the light. Its homeostasis has been reported in several studies to be affected by synthesis, transport from buds, or regulation of other plant hormones ( Gomez-Roldan et al. , 2008 ; Ferguson and Beveridge, 2009 ; Domagalska and Leyser, 2011 ; Dun et al. , 2012 ). Auxin flow from the shoot tip is suggested to prevent auxin flow from the axillary buds or, in an alternative theory, to inhibit CK and/or promote SL synthesis in the axillary buds ( Brewer et al. , 2009 , 2013 ; Agusti et al. , 2011 ; Dun et al. , 2012 ). The role of GA in AD is unclear: in some plant systems, it seems to play an inhibitory role in LT outgrowth ( Scott et al. , 1967 ; Koorneef et al. , 1985 ; Mauriat et al. , 2011 ); in others, it has been shown to stimulate LT development ( Marth et al. , 1956 ; Elfving et al. , 2011 ; Ni et al. , 2015 ). The involvement of GA biosynthesis in the effects of light or short-day photoperiods on bud burst has also been demonstrated in plants ( Choubane et al. , 2012 ; Zawaski and Busov, 2014 ). Since etiolated sprouting of potato tubers was investigated as a case study for storage-organ branching under dark conditions, the key hormones responsible for bud burst and development and their interaction were determined.

As shown previously by Hartmann et al. (2011) , 6-BA and GA ₃ induced burst and elongation, respectively, of detached potato buds ( Fig. 2 ). Application of NAA eliminated these effects on detached buds ( Fig. 2 ). A similar effect could be found in whole tubers, when phytohormones were added to the AP or to the fifth LT, suggesting that the AP is a strong sink for tuber phytohormones ( Fig. 3 ). Indeed, GA ₃ was detected at higher levels in the AP 24h after its application to the LT ( Fig. 4 ). Although the GA ₃ level was enhanced in the LT as well, only the AP elongated, suggesting dose-dependence or the existence of other factors that need to be combined to induce early bud growth. In a study of transgenic lines with ectopic expression of GA-biosynthetic genes, Hartmann et al. (2011) speculated that the amount of bioactive GA must exceed a certain limit in responding cells to stimulate sprout outgrowth.

Not surprising is the fact that among auxins, only free IAA was detected in the tuber buds ( Supplementary Table S1 and Fig. 4 ), as it is considered a major player in potato tuber dormancy release and sprouting ( Sorce et al. , 2000 , 2009 ). Indeed, the higher initial level of IAA in the LT can be explained by its late sprouting expectation. The equalization pattern of IAA concentration between the LT and AP is more difficult to explain, but it may suggest local transport, from the buds to the parenchyma, as a result of GA ₃ treatment.

The major increase in the Cis-ZOG level only in the AP ( Fig. 4 and Supplementary Table S1 ) suggests its involvement with the early events leading to bud elongation. Other CK derivatives such as (trans) zeatin-O-glucoside, t-Z, c-Z, dihydrozeatin, (trans) zeatin riboside, (cis) zeatin riboside, dihydrozeatin riboside, isopentenyladenine, and isopentenyladenosine were below the limit of quantification/detection. In a previous study, c-Z was proposed as a possible regulator of potato tuber dormancy ( Suttle and Banowetz, 2000 ); however, no clear correlation was found between expression of CK oxidase/dehydrogenase-like genes and progression of tuber dormancy ( Suttle et al. , 2014 ). This finding was compatible with the observation that the synthetic phenylurea and nitroguanidine CKs, which cannot be inactivated by endogenous CKX, are more effective in dormancy termination than the naturally occurring zeatin species ( Suttle, 2008 ). O-glucosyltransferase activity is specific to c-Z, and is considered to be important in CK transport, storage, and protection against CKXs ( Martin et al. , 1999 , 2001 ; Veach et al. , 2003 ). In potato tubers, O-glycosylation of c-Z specifically in the AP, induced by GA treatment, might be involved in the function of c-Z in determining bud growth priority.

Hormone transport in the peripheral vascular system

The dual-bud tissue-strip assay indicated that 6-BA, GA ₃ , and NAA transport can occur in the peripheral part of the tuber, probably via the vascular ring ( Fig. 5 ). Hormonal regulation of LT outgrowth involves the antagonistic actions of auxin, CK, and SL ( Beveridge and Kyozuka, 2010 ; Domagalska and Leyser, 2011 ; Müller and Leyser, 2011 ). It is well accepted that auxin synthesized in the shoot apex moves basipetally in the polar auxin transport stream and that CK travels acropetally through the xylem into the axillary bud, directly promoting its outgrowth (reviewed by Durbak et al. , 2012 ). In the whole potato tuber, as well as in the tissue-strip assay, the AP was the strongest sink for the phytohormones. Its continuing function in the strip suggests its autonomous strength, with no need for whole-tuber functions ( Figs 5 , 6 ). The ability of exogenous NAA to eliminate the growth induction by exogenous 6-BA or GA ₃ was shown in the tissue strip as well, in both cases – treatment of the AP or LT – and it inhibited the AP. Uptake of the synthetic auxin NAA is known to be independent of auxin-influx carriers and appears to be a consequence of perturbed auxin distribution affecting auxin signaling ( Delbarre et al. , 1996 ). These data identify a novel role for auxin import into the AP during bud burst. NAA is not recognized by the auxin importers AUX1/Lax1 ( Kramer and Bennett, 2006 ; Swarup et al. , 2008 ) and is probably transported via simple flux to the growing AP. Use of the triple-bud strips emphasized the strength of the AP sink ( Fig. 6 ). The reduced effect of GA ₃ on the LTs, even when the hormone is forced to move through their base on its way to the AP, suggests that distance from the site of application is not a significant factor in its action. GAs have been shown to influence source–sink relationships in various plant processes. One way they function is by having a strong influence on phloem loading and regulation of sucrose synthesis, the major end product of photosynthetic carbon metabolism (reviewed by Iqbal et al. , 2011 ). Recently Mason et al. (2014) suggested that AD in the model pea plant is predominantly controlled by the shoot tip’s intense demand for sugars, which limits sugar availability to the LTs. In the tuber system, the AP’s specific GA ₃ sink may be involved in GA’s crucial role in bud activation via establishment of a stronger sink that eventually leads to AD.

Expression of hormone-related genes

Our transcriptomic analysis provides the first expressed-gene dataset for comparison of potato tuber AP and LTs in response to GA ₃ stimulation. In the AP, which acts as a stronger sink for GA ₃ , up-regulation was observed for genes related to cell cycle, proliferation, and increased metabolism relative to the LT. On the other hand, genes associated with processes such as protein degradation, stress, and defense responses were observed to be up-regulated in the LT ( Fig. 8 ). These initial differences in gene expression may be associated with the AP being a stronger sink for nutrients and, later, with its dominance over the LTs.

The role of phytohormones in plant AD in general, and in potatoes in particular, is not completely clear. In the present study, the expression of hormone-related genes was analyzed by comparing the AP and LT transcriptomes, before and up to 72h after GA ₃ treatment. In other plants, GA has been shown to interact with other phytohormones via diverse mechanisms, which act at both the hormone-response and hormone-biosynthesis levels (reviewed by Weiss and Ori, 2007 ).

Comparative analysis of the LT and AP transcriptomes revealed higher expression of ABA-responsive genes, such as ABF , PBR , and HVA22 (details of genes on heat map in Supplementary Table S2 ), and ABA metabolism-related genes, such as NCED 3 , 4 , and 6 , in the LT, suggesting a role for ABA in inhibiting LT growth, especially at time 0 ( Destefano-Beltran et al. , 2006 ). Down-regulation of some of these genes 24 or 72h after GA ₃ treatment ( Fig. 9 ) might be one of the mechanisms of bud-burst induction in the LT at later stages, which is not needed by the AP as it is more developed and in the elongation stage. These results correlate well with the faster AP response to GA ₃ .

There are several reports suggesting that the CK level increases in the bud as part of dormancy break and bud burst ( Turnbull and Hanke, 1985 ; Suttle and Banowetz, 2000 ). The sensitivity of the tuber buds to exogenous CK during dormancy break increases gradually ( Turnbull and Hanke, 1985 ; Suttle, 2001 ). CK receptor homologs, such as AHK , and CK metabolism genes, such as CKX s and ITP s, showed elevated expression in the LT and AP 24 and 72h after GA ₃ application ( Fig. 9 ). Nevertheless, based on our experiments and a previous study ( Hartmann et al. , 2011 ), CK is expected to be involved in bud burst. After GA ₃ application to the LT, the AP quickly moves to the elongation stage with a concomitant increase in cis-ZOG ( Fig. 4D ), and most of the CK metabolism probably moves to the LTs.

Sprouting induction by GAs, especially in the shallow dormancy period of potato tubers, has been reported in several studies ( Brian et al. , 1955 ; Rappaport et al. , 1957 ; Hemberg, 1985 ; Suttle, 2004 a ). Detectable levels of endogenous GAs were only found when sprouting was already observed ( Suttle, 2004 a ). Although only GA ₃ , which was applied exogenously, was detected in the AP ( Fig. 4 and Supplementary Table S1 ), transcriptome analysis showed up-regulation of GA-metabolism and GA-responsive genes mostly in the AP ( Fig. 9 ). The AP’s stronger sink ( Fig. 4 ) probably caused the observed higher expression of these GA-responsive and -metabolism genes as part of the AP’s rapid elongation ( Fig. 9 ).

Ethylene is known as a growth-inhibiting hormone that affects GA-mediated DELLA degradation ( Achard et al. , 2003 ; Dugardeyn et al. , 2008 ). Ethylene has been shown to play a role in the regulation of tuber dormancy ( Suttle, 1998 , 2007 ). Most ethylene-related genes were down-regulated in response to GA ₃ , especially in the AP ( Fig. 9 ). Suttle (2009) suggested that ethylene is not involved in dormancy break. In our study, some of the ethylene-metabolism genes were down-regulated in the AP and some of the ethylene receptor-encoding genes were up-regulated in the LT, both after GA ₃ treatment ( Fig. 9 ), suggesting ethylene’s role in controlling bud burst and elongation.

The up-regulation of auxin transport genes, such as PIN 2 , 3 , 5 , and 6 , emphasizes the interaction of the AP with GA ₃ ( Fig. 9 ). An interaction between GA and auxin signaling is suggested to be part of the predominant AP response. Higher GA levels on the lower side of the root have been shown to correlate with increasing amounts of PIN2 at the plasma membrane during Arabidopsis root gravitropism ( Löfke et al. , 2013 ). In general, the transcriptome analysis supported the involvement of phytohormones in the predominant effect of GA ₃ on the AP.

Overall, a stronger sink demand for GA is suggested to support the dominance of the AP in etiolated growth. AP development, induced by GA ₃ , interacts with auxin signaling to induce predominant AP elongation, resulting in AD.

Supplementary data

Supplementary data are available at JXB online.

Table S1. Endogenous levels of phytohormones before (0) and 24 or 72h after GA ₃ treatment

Table S2. Detailed list of differentially expressed genes related to hormonal regulation.

Fig. S1. Venn diagram comparison of differentially expressed genes after GA ₃ treatment applied to the fifth-position lateral bud.

Fig. S2. Gene ontology enrichment analysis for clusters found in Fig. 8 .

Acknowledgements

We thank Dr Einat Zelinger from The Hebrew University of Jerusalem, Robert H. Smith Faculty of Agriculture, Food and Environment, Rehovot, Israel for her assistance with the cryostat, and confocal and image analyses.

Author notes

Editor: Robert Hancock, The James Hutton Institute