The SHORT-ROOT -like gene PtSHR2B is involved in Populus phellogen activity

SHORT-ROOT (SHR) is a GRAS transcription factor first characterized for its role in the specification of the stem cell niche and radial patterning in Arabidopsis thaliana ( At ) roots. Three SHR -like genes have been identified in Populus trichocarpa ( Pt ). PtSHR1 shares high similarity with AtSHR over the entire length of the coding sequence. The two other Populus SHR -like genes, PtSHR2A and PtSHR2B , are shorter in their 5' ends when compared with AtSHR . Unlike PtSHR1 , that is expressed throughout the cambial zone of greenhouse-grown Populus trees, PtSHR2Bprom:uidA expression was detected in the phellogen. Additionally, PtSHR1 and PtSHR2B expression patterns markedly differ in the shoot apex and roots of in vitro plants. Transgenic hybrid aspen expressing PtSHR2B under the 35S constitutive promoter showed overall reduced tree growth while the proportion of bark increased relative to the wood. Reverse transcription–quantitative PCR (RT–qPCR) revealed increased transcript levels of cytokinin metabolism and response-related genes in the transgenic plants consistent with an increase of total cytokinin levels. This was confirmed by cytokinin quantification by LC-MS/MS. Our results indicate that PtSHR2B appears to function in the phellogen and therefore in the regulation of phellem and periderm formation, possibly acting through modulation of cytokinin homeostasis. Furthermore, this work points to a functional diversification of SHR after the divergence of the Populus and Arabidopsis lineages. This finding may contribute to selection and breeding strategies of cork oak in which, unlike Populus , the phellogen is active throughout the entire tree lifespan, being at the basis of a highly profitable cork industry.


Introduction
Plant growth occurs from specialized regions called meristems. Mitotic divisions in the meristems produce the cells that eventually differentiate into the organs and tissues that comprise the body of the plant. Four main meristems exist in woody perennials, the root and shoot apical meristems, that provide cells for shoot and root growth, respectively, and the vascular cambium and cork cambium (phellogen) that generate cells for the secondary, or radial, growth of the stem, branches, and roots.
During secondary growth, the epidermis is replaced in the stems and roots by the periderm. This protective tissue is derived from the activity of the phellogen or cork cambium that forms a continuous ring of meristematic cells around the stem and roots. Periclinal divisions of the phellogen initials give rise to cells that differentiate into phelloderm towards the inside of the stem and into phellem towards the outside (Evert, 2006), in a way similar to the differentiation of vascular tissues from the vascular cambium. In a few species, of which the cork oak is an extreme example, the activity of the phellogen contributes to a significant enlargement of the trunk, producing a thick layer of phellem or cork. In adult cork oak trees, cork is regularly stripped from the tree and used for several industrial applications due to its exceptional impermeability, insulation, density, high energy absorption, resilience, and elasticity properties (Silva et al., 2005).
The SHR transcription factor belongs to the GRAS family of plant-specific proteins that are characterized by a variable N-terminal domain but a highly conserved C-terminal domain (Helariutta et al., 2000;Bolle, 2004). The Arabidopsis SHR (AtSHR) has been well characterized, being a key regulator, along with the related GRAS protein SCARECROW (SCR), of radial patterning and stem cell niche specification in the roots (Benfey et al., 1993;Di Laurenzio et al., 1996;Helariutta et al., 2000;Nakajima et al., 2001). It is essential for the asymmetric cell divisions of the cortex/endodermal initial (CEI) (Benfey et al., 1993;Helariutta et al., 2000;Nakajima et al., 2001), and for the periclinal divisions of cortex cells in a maturing root (Paquette and Benfey, 2005). Fukaki et al. (1998) demonstrated that AtSHR is also involved in radial patterning in the shoot. More recently, Dhondt et al. (2010) demonstrated that, similarly to the Arabidopsis root, SHR functions in association with its downstream target, SCR, in the regulation of cell proliferation and vascular differentiation in leaves. The SHR/SCR mechanism therefore appears to have been co-opted to regulate cell proliferation and differentiation in multiple organs. SHR modulates the expression of genes involved in a wide range of processes during Arabidopsis root development, including transcriptional regulation, signalling, and response to hormones, and in the regulation of cell cycle genes (Levesque et al., 2006;Sozzani et al., 2010). In Arabidopsis, correct patterning of the central vascular cylinder is mediated through movement of the SHR protein from the stele into the endodermis (Nakajima et al., 2001), where it activates its target, SCR, that together activate miR165a and miR166b (Carlsbecker et al., 2010). The regulation of vascular patterning by SHR in the Arabidopsis root involves the modulation of cytokinin (CK) homeostasis through the direct regulation of the cytokinin-degrading enzyme, CYTOKININ OXIDASE 3 (CKX3) (Cui et al., 2011;Hao and Cui, 2012).
SHR has also been studied in tree species such as Pinus radiata, where it was suggested to have roles in root meristem formation and maintenance, and in the cambial region of hypocotyls (Solé et al., 2008). The putative Populus orthologue of AtSHR, PtSHR1, is expressed in the cambial zone (Schrader et al., 2004;Wang et al., 2011) and functions as a regulator of cell division and meristem activity in the shoots (Wang et al., 2011). Partial suppression of the PtSHR1 transcript in transgenic lines leads to taller trees with a larger vascular cambium due to an increase in cell proliferation in the cambial zone (Wang et al., 2011). In both Arabidopsis and Populus, it has been suggested that SHR regulates growth through the control of cell divisions, in a concentrationdependent manner (Paquette and Benfey, 2005;Wang et al., 2011;Koizumi et al., 2012).
Whereas there is only one SHR gene in Arabidopsis, three SHR-like genes have been identified in the Populus genome, PtSHR1, PtSHR2A, and PtSHR2B (Wang et al., 2011). Based on sequence similarity and on functional studies with the PtSHR1 coding sequence, driven by the AtSHR promoter, PtSHR1 is considered to be the putative ortholog of the Arabidopsis SHR (Wang et al., 2011). In this study, PtSHR2B was investigated in order to better understand its role in meristem function in hybrid aspen. We show that PtSHR2B is expressed in the phellogen, pointing to a regulatory role in this meristem during secondary growth. Overexpression of PtSHR2B in hybrid aspen not only affected overall plant growth, but altered the ratio between the amount of wood and bark tissues in the stem. We further present experiments that indicate that PtSHR2B may act, at least partially, through the regulation of CK homeostasis.

Plant material
Hybrid aspen (Populus tremula L.×Populus tremuloides Michx.; Clone T89) was propagated in vitro on half-strength basal MS medium (Murashige and Skoog, 1962) and maintained in a growth chamber at 21 ºC in a 16 h light/8 h dark photoperiod. For greenhouse experiments, in vitro established wild-type and transgenic plants were placed in a soil:peat:perlite (1:3:1) potting mix and acclimatized in a growth chamber, gradually decreasing the humidity from 95% to 70% over 5 weeks before transfer to the greenhouse, where plants were grown for a minimum of 10 weeks prior to analysis. The position of all pots within the greenhouse was changed weekly to minimize positional bias experimental error.
After 10 weeks in the greenhouse, the five youngest fully expanded leaves were collected (Supplementary Fig. S1A at JXB online). Bark was isolated by peeling off the stem across the cambial zone. The remaining stem tissues (xylem and pith), hereafter termed 'wood', were collected together. All samples were immediately frozen in liquid nitrogen and stored at -80 ºC until further processing. Intact stem sections were also collected and fixed in FAA (5% formaldehyde, 5% acetic acid, 50% ethanol) for anatomical analysis.
Histochemical GUS assay GUS assays were performed in transformed hybrid aspen leaves, roots, and shoot apex collected from 6-week-old in vitro grown shoots, and 6-month-old and 1-year-old stems of greenhouse-grown trees. Tissues were placed in ice-cold 90% acetone for 30 min and then washed in water prior to immersion in the GUS staining solution [10 mM sodium phosphate buffer pH 7.0, 0.5% Triton X-100, 2 mM potassium ferricyanide, and 2 mM X-Gluc (5-bromo-4chloro-3-indolyl β-d-glucuronide)], vacuum infiltrated, and incubated overnight in the dark at 37 ºC. After washing in water, tested leaves, roots, and shoot apex of in vitro plants were gradually dehydrated to 70% ethanol and stem sections of the greenhouse-grown plants were fixed in FAA and then included in Technovit 7100 resin (Heraeus Kulzer), according to the manufacturer`s instructions, with minor modifications: after vacuum infiltration, samples were left for 2 d at 4 ºC in the pre-infiltration solution. The solution was then replaced and samples left for another 7 d at 4 ºC. The material was subsequently placed in the infiltration solution and left for 1-3 weeks at 4 ºC, followed by polymerization at room temperature.
Stereomicroscope observations were performed with a Nikon SMZ800, and images were captured using an Olympus SC30 camera and software. Microscope observations were made with a Nikon Inverted Microscope Eclipse TE300, and images taken with a Nikon DS-Fi1 camera using the NIS-Elements F3.0 software.

Anatomical analysis and growth measurements
Stem pieces of 1-and 2-year-old shoots of wild-type hybrid aspen were collected and fixed in ice-cold FAA, as previously described, vacuum infiltrated, and left overnight in a desiccator at 4 ºC. After gradual dehydration to 100% ethanol, tissues were embedded in resin as described above, and 6-8 µm thick sections were stained with Toluidine Blue O. Several growth parameters were analysed in 10-week-old greenhouse-grown 35S:PtSHR2B trees ( Supplementary Fig. S1A). Tree height and total length between the 10th (EN10) and the 17th (EN17) internodes from the shoot tip were recorded. Stem diameters were measured at the reference internode (EN14), showing fully differentiated secondary vascular tissues and chosen for comparison between transgenic and wild-type plants, and at the stem base (ENbase), corresponding to an internode between the 20th and 25th internodes depending on the tree. Measurements of distances between different stem tissues were taken at a minimum of four positions around the circumference of the stem sections using ImageJ software (Abràmoff et al., 2004;Schneider et al., 2012) (Supplementary Fig. S1B). The total lamina area of the five leaves surrounding EN14 was determined using a leaf area meter (LI-3000A, LI-COR Inc.).

Reverse transcription-quantitative PCR (RT-qPCR)
The first five fully expanded leaves down from the shoot tip, and bark and wood from the ENbase were collected and ground to a fine powder ( Supplementary Fig. S1A) using either a mortar and pestle (leaves and bark) or a grinder mill (M 20 Universal mill, Ika). Isolation of bark and wood tissues was done as described in the 'Plant material' section. Total RNA from leaves and bark was isolated as described by Reid et al. (2006) with minor modifications (Marum et al., 2012). Total RNA from wood samples was extracted using the protocol as described in Chang et al., (1993), and all RNA samples were treated with TURBO DNase (Ambion) according to the manufacturer`s instructions. cDNA synthesis was performed from 1.5 µg of DNase-treated RNA, using a Transcriptor High Fidelity cDNA Synthesis Kit (Roche) with anchored oligo(dT) 18 primers. Quantitative real-time PCR (qPCR) was carried out in 96-well plates in a LightCycler 480 (Roche) using SYBR Green I Master Mix (Roche). Primers for amplifying a transcript fragment of PtSHR2B, 5'F-CAGCAATACCCTTTGCACACAG-3' and 5'R-ACCCAGTCCTTCCTTTGTG-3', were designed using the P. trichocarpa genome version 1.1 (http://genome.jgipsf.org/Poptr1_1/Poptr1_1.home.html) and the gene sequence (eugene3.00640143) described by Wang et al. (2011). For amplification of PtCKX3 (Potri.006G152500.1) transcripts, specific primers 5'F-TCAGATCCAAACCCTTGATTTC-3' and 5'R-CAGTAAAAGGGGTGTAGTT-3' were designed using P. trichocarpa genome version 3, from Phytozome (http:// www.phytozome.net/). For amplification of PtRR7 transcripts (Potri.016G038000.1), primers were as previously described (Nieminen et al., 2008). PtCYP2 (Potri.004G168800.1) was used as a reference gene (Brunner et al., 2004;Milhinhos et al., 2013). The PCR program used was 95 ºC for 10 min, 45 cycles of 10 s at 95 ºC, 20 s at 60 ºC for PtRR7 and PtCKX3 or 20 s at 63 ºC for PtSHR2B, and 10 s at 72 ºC. The annealing temperature for the reference gene primers was 60 ºC or 63 ºC, depending on the experiment. Three technical replicates were used for each of the three biological samples in each experiment. To normalize values obtained from different plates, a calibrator sample consisting of cDNA synthesized from RNA from leaves of a transgenic line was used in each plate. Normalized relative quantities were obtained through the ΔΔC T method (Livak and Schmittgen, 2001;Pfaffl, 2001;Hellemans et al., 2007) and the amplification efficiency determined using Real-Time PCR Miner (Zhao and Fernald, 2005).

Quantification of cytokinins
To quantify the CK levels in bark and wood tissues, samples were ground into a fine powder. CKs were extracted and isolated from ~10 mg of frozen tissues from bark and wood as previously described (Svačinová et al., 2012), including modifications described by Antoniadi et al. (2015). To each extract, the stable isotope-labelled CK internal standards (0.1 pmol of CK bases, ribosides, N-glucosides, 0.25 pmol of O-glucosides, and 0.5 pmol of nucleotides) were further added as a reference. Purified samples were analysed by an LC-MS/MS system consisting of an ACQUITY UPLC ® System (Waters) and a Xevo™ TQ-S (Waters) triple quadrupole mass spectrometer. Quantification was obtained using a multiple reaction monitoring (MRM) mode of selected precursor ions and the appropriate product ion. Five independent biological replicates were analysed for each sample.

Statistical analysis
The assessment of statistical significance in transcript profiles and phenotypic parameters was carried out using non-parametric analysis, Mann-Whitney U-test. A significance level of P=0.05 was used. An ANOVA was performed to assess the statistical significance in the quantification of CKs. Statistics were performed using the Statistica (StatSoft Inc., http://www.statsoft.com) software package.

Results
In this work, the Populus SHR-like gene PtSHR2B was characterized and compared with the putative Populus orthologue of the Arabidopsis SHR gene, PtSHR1 (Wang et al., 2011). AtSHR and PtSHR1 genes have been previously implicated in the regulation of primary apical meristems and vascular cambium activity. Transverse sections of hybrid aspen stems at different developmental stages showed a distinct phellogen meristematic layer, already present at the end of the first year of growth and characterized by rectangular cells that are flattened radially and divide mostly by periclinal division ( Supplementary Fig. S1C). The characteristic layer of suberized phellem cells in the periderm could be observed. The periderm in 2-year-old stems had only a slight increase in phellem layer thickness compared with 1-year-old stems ( Supplementary Fig. S1C, D).

PtSHR1 and PtSHR2B show different expression patterns
While tissues from wild-type controls were always negative to GUS histochemical assay (Fig. 1A-D), analysis of hybrid aspen plants carrying either the PtSHR1 or PtSHR2B promoter driving uidA expression (PtSHR n prom:uidA) indicated different patterns of promoter activity. Analysis of PtSHR1 promoter activity in greenhouse-grown trees that had undergone substantial secondary growth showed GUS throughout the cambial zone and in xylem rays (Fig. 1E, F). Additionally, GUS expression was found in the leaf vasculature and in the root stele of in vitro plants (Fig. 1G,  H), similarly to its homologue in Arabidopsis (AtSHR) (Helariutta et al., 2000;Dhondt et al., 2010;Wang et al., 2011). In the shoot apex, GUS staining was observed in the apical meristem and vasculature (Fig. 1I). In the case of the PtSHR2B promoter, the greenhouse-grown trees showed GUS staining strongly localized in the phellogen cell layer (Fig. 1J, K), suggesting a specific function for the modified version of SHR in this meristem. In leaves from in vitro plants, GUS expression was not detected (Fig. 1L), but in the roots GUS staining showed a stark contrast to that observed with the PtSHR1 promoter, with GUS expression being observed at the root tip (Fig. 1M). Differences between the expression driven by each of the two promoters were also found in in vitro developing shoot tip where PtSHR2Bprom:uidA staining was restricted to apical and axillary meristems (Fig. 1N). Since no GUS signal was ever detected in any of the analysed tissues from plants carrying the PtSHR2Aprom:uidA construct, the study only proceeded with the analysis of PtSHR2B.
Profiling of PtSHR2B transcript levels by RT-qPCR in the tissues of wild-type greenhouse-grown hybrid aspen revealed the highest levels in the bark, with significantly lower levels in wood and leaf tissues (Fig. 2). These results corroborate the GUS staining observations, confirming that expression is predominantly in the phellogen, although not restricted to it (Figs 1J, K, 2).

Ectopic expression of PtSHR2B reduces overall tree growth
To explore the role of PtSHR2B in hybrid aspen, the PtSHR2B coding region was isolated and transformed into hybrid aspen under the control of the constitutive CaMV 35S promoter. RT-qPCR confirmed the ectopic expression of PtSHR2B in leaves and stem tissues of in vitro 35S:PtSHR2B plants (Fig. 3A). However, in the greenhouse-grown trees, the increased accumulation of PtSHR2B transcript was observed in leaves and bark, but not in wood tissues (Fig. 3B). No obvious phenotype could be observed in in vitro grown plants. However, 10-week old greenhousegrown transgenic trees showed a reduced growth compared with the wild type ( Fig. 4; Supplementary Fig. S2). Tree height in all of the transgenic lines was significantly reduced when compared with the wild-type plants (Fig. 4A, B). Both control and transgenic trees experienced growth deceleration between weeks 7 and 10, possibly due to a greenhouse temperature increase during this period. Stem diameter was significantly reduced compared with the wild type at the base of the transgenic trees (Fig. 4C) although no change could be observed in the reference internode (EN14). The reduction in height in the transgenic trees was primarily the result of a reduced internode length, as the mean internode length between EN10 and EN17 was significantly shorter in the transgenic trees compared with the wild type (Fig. 4D). The total number of internodes was slightly reduced, but this was only significant in one of the transgenic lines (Fig. 4E). The total average fresh weight and lamina area of the five leaves surrounding the reference internode was also significantly reduced in the transgenic trees (Fig. 4F, G).

Altered stem anatomy in 35S:PtSHR2B ectopic expression plants
Transverse sections taken from stems of the 35S:PtSHR2B transgenic trees were analysed by light microscopy. The secondary xylem [radial distance from the pith side of the lignified xylem to the cambial zone ( Supplementary Fig.  S1B)] was significantly reduced in the transgenic trees compared with the wild type in two of the three transgenic lines (Fig. 5A), whereas the bark layer was wider than in the wild type (Fig. 5A). The difference in the proportion of wood and bark tissues was more evident at the base of the stem, where secondary growth is far more extensive (Fig. 5B). Additionally, the ratio between the phellem and the stem radius was reduced in two out of the three transgenic lines (Fig. 5C, D). Some variation was seen between the independent transformant lines, which is commonly observed and consistent with previous work using Populus Milhinhos et al., 2013).

Genes involved in cytokinin metabolism and response are altered in transgenic plants
CK has been linked to SHR function in Arabidopsis roots where SHR controls vascular patterning through its effect on CK homeostasis (Cui et al., 2011). We analysed the transcript levels of a central CK primary response gene, the A-type response regulator PtRR7 (Nieminen et al., 2008;Ramírez-Carvajal et al., 2008), in bark and wood tissues of the 35S:PtSHR2B transgenic trees. PtRR7 transcript levels were higher in the bark of all PtSHR2B overexpression lines compared with the wild type (Fig. 6A), indicating altered CK signalling in this tissue.
In Arabidopsis roots, SHR is known to act, at least partially, by directly regulating the expression of CYTOKININ OXIDASE 3, AtCKX3 (Cui et al., 2011). We also analysed transcript levels for the putative Populus AtCKX3 orthologue, PtCKX3, in bark and wood tissues of the transgenic plants. Transcripts for PtCKX3 were more abundant in the bark of the transgenic trees, compared with the wild type (Fig. 6B). Increased transcript levels for these genes in the bark indicate that PtSHR2B levels are important for the regulation of CK metabolism in this tissue. Transcript levels for both genes were also increased in wood tissue (Fig. 6A, B), but to a lesser extent than in bark and only in some transgenic lines.

Quantification of cytokinins in stem tissues
To clarify the link between PtSHR2B and CK homeostasis, we quantified the levels of CKs present in wood and bark tissues of the transgenic trees. Naturally occurring CKs are adenine derivatives with either isoprenoid or aromatic side chains. Isoprenoid CKs can be distinguished as isopentenyladenine (iP)-, trans-zeatin (tZ)-, cis-zeatin (cZ)-, or dihydrozeatin (DHZ)-type derivatives, depending on the hydroxylation and reduction of the side chain (Ha et al., 2012). Total CK levels were altered in both tissues, as was the distribution of its derivatives (Fig. 7; Table 1). All 35S:PtSHR2B transgenic trees presented high levels of total CKs in bark but lower levels in wood tissues (Fig. 7).

Discussion
The vascular cambium functioning and the development of vascular tissues have been the subject of numerous reports (Baucher et al., 2007;Du and Groover, 2010;Schuetz et al., 2012;Sanchez et al., 2012). In contrast, despite the important role that the phellogen plays in providing cells for the development of the protective layer on stems, branches, and roots, it has received minimal attention. Under our growth conditions, a surrounding meristematic phellogen and a phellem (cork) layer had formed in 1-year-old hybrid aspen stems. The thickness of this layer was almost unchanged between 1and 2-year-old hybrid aspen stems. Consistency in the phellem layer thickness during the lifespan of the close relative Populus tremuloides Michx. has long been known (Kaufert, 1937), indicating that in Populus species phellogen cell divisions are matched by a shedding of cork layer cells.
Populus is recognized as a model species in the study of angiosperm tree function. It has a high annual rate of secondary growth, and like other model species has the advantages of being relatively easily genetically modified and cultured. There are also now ample molecular, genomic, and bioinformatics resources available for various Populus species. Importantly, the P. trichocarpa genome has been fully sequenced (Tuskan et al., 2006). The sequence data indicate that two whole-genome duplication events occurred in ancestors of the species (Tuskan et al., 2006). Although duplicated genes are often lost over evolutionary time, higher gene retention is often found, particularly for specific classes of genes, such as: (i) genes with regulatory functions, namely transcription factors and developmental regulators (Blanc and Wolfe, 2004;Seoighe and Gehring, 2004;Carretero-Paulet and Fares, 2012); and (ii) for genes derived from a previous round of genome duplication (Seoighe and Gehring, 2004). The biased retention is most probably because multiple copies of the retained genes impart specific beneficial effects for the organism (Seoighe and Gehring, 2004;Carretero-Paulet and Fares, 2012). One of the ways in which multiple genes can be beneficial is through speciation, leading to divergent function and expression patterns (Blanc and Wolfe, 2004;Tuskan et al., 2006;Rodgers-Melnick et al., 2012). In Populus, it has been hypothesized that after duplication, gene preservation is influenced by a combination of subfunctionalization and . The presence of three SHR-like genes in the Populus genome indicates that SHR fits these criteria. In this work we show that the patterns of GUS expression driven by Populus PtSHR1 and PtSHR2B promoters differ markedly both in the shoot apex and in the roots. Commensurate with AtSHR expression in Arabidopsis, GUS staining in the roots of the PtSHR1prom:uidA plants was confined to the stele. In contrast, the PtSHR2B promoter activity was detected in the root tip, suggesting that the two genes have different functions in the root. In greenhouse-grown plants with extensive secondary growth, GUS expression driven by both PtSHR1 and PtSHR2B promoters was strongly associated with the lateral meristems. However, while GUS staining was detected in the vascular cambium of the PtSHR1prom:uidA plants, the PtSHR2B promoter drove GUS expression in the phellogen. PtSHR1 has previously been reported to be expressed in the vascular cambium, and to regulate its activity (Schrader et al., 2004;Wang et al., 2011). Furthermore, AtSHR and PtSHR1 have been shown to have broad activity in meristems in the roots and shoots (Schrader et al., 2004;Wang et al., 2011). However, this is the first work reporting an SHR-like gene promoter activity in the phellogen, suggesting that PtSHR2B fulfils an important function in this lateral meristem. This finding led us to explore further the function of PtSHR2B by ectopically expressing it in hybrid aspen. Compared with the wild type, overall growth was reduced in transgenic trees as evaluated by several parameters to assess primary and secondary growth. Total tree height and mean internode length between EN10 and EN17 were amongst the most significantly reduced growth parameters in all the transformed lines. The total number of internodes was only slightly reduced, further indicating that stem elongation was inhibited in the transgenic lines. The stems of the transgenic trees were less tapered than those of the wild-type plants since both groups had similar stem diameters in the upper region of the stem (at EN14), but the stems of the transgenics were appreciably narrower at the base (EN20-EN25). When the stem tissues were analysed in more detail, we verified a relative reduction in the wood layer, while an increase in the proportion of bark, composed of phloem, phelloderm, and phellem, was observed. We could not correlate the increase in the bark width with an increase of the phellem given that the shedding of cork layer cells in Populus occurs simultaneously with cell division of phellogen initials (Kaufert, 1937). Altogether, the phenotypic analysis of these plants suggests that, in addition to the effect on secondary growth patterns, PtSHR2B ectopic expression affects overall plant growth. When studying the function of PtSHR1 in hybrid aspen, Wang et al. (2011) have shown that its partial suppression enhanced overall plant growth, and suggested that the protein acts as a dose-dependent negative regulator of meristem activity. Our results do not allow a conclusion to be drawn on the dose-dependent action of PtSHR2B but support that it might interfere in plant growth through similar mechanisms to those of PtSHR1.
In order to investigate the possible mechanisms by which PtSHR2B may exert its function, we have analysed the expression of the CK-related genes PtRR7 and PtCKX3, which encode a CK primary response and a CK degradation protein, respectively. As inducers of cell division and differentiation, CKs are central regulators of plant development (Miyawaki et al., 2006;Sakakibara, 2006). In fact, the importance of CK signalling in the regulation of cambial activity in Populus stems has been previously demonstrated (Nieminen et al., 2008). In our work, we found that the expression of both genes was significantly up-regulated in the bark tissues of all the transgenic lines, but their expression levels were either similar or only slightly increased in the wood tissues when compared with the wild type. PtRR7 transcript levels were reported to be positively correlated with the amount of CK present across the stems of hybrid aspen trees (Nieminen et al., 2008). Cytokinin oxidase gene expression has also been shown to be up-regulated in response to increased CK levels (Motyka et al., 1996;Jiao et al., 2003). Additionally, in Arabidopsis, the CKX3 transcript levels were significantly reduced in the shr roots (Cui et al., 2011). Our transcript quantification results would seem consistent with increased CK endogenous levels in the tissues where expression is higher. In fact, through quantification of CKs we were able to confirm that endogenous total CK levels were significantly higher in the bark tissues, in two out of the three transgenic lines. This trend was not verified in wood tissues that, despite the lower CK levels in two out of three lines, showed a slight up-regulation of PtRR7. Differences in CK biosynthesis and homeostasis regulation, or even their crosstalk with auxin in the vascular tissues may contribute to explain this observation (Kieber and Schaller, 2014), although additional studies would be required to clarify this issue.
The mechanisms by which the overexpression of PtSHR2B might lead to increased CK levels in the bark tissues were not addressed in our study. However, it is tempting to speculate that similarly to AtSHR (Sebastian et al., 2015), PtSHR2B may interfere with CK homeostasis. It is possible that the increased CK levels in the bark (containing the phellogen layer) result from an increased amount of PtSHR2B transcripts which may regulate target genes with specific domains of expression in these tissues. The decrease in the wood layer, indicative of modified cambium activity, might also be explained by the slightly reduced levels of total CKs in the wood tissues. This will need to be tested in future studies with tissue-specific promoters, since secondary effects of ectopic expression of PtSHR2B cannot be ruled out.

Supplementary data
Supplementary data are available at JXB online. Figure S1. Schematic representation of tissue sampling, stem section measurements, and anatomical aspects of hybrid aspen stem. Figure S2. Growth curves of the wild-type and 35S:PtSHR2B hybrid aspen plants.