Peptide encoding Populus CLV3/ESR-RELATED 47 (PttCLE47) promotes cambial development and secondary xylem formation in hybrid aspen.

The CLAVATA3 (CLV3)/EMBRYO SURROUNDING REGION (ESR)-RELATED (CLE) peptide ligands in connection with their receptors are important players in cell-to-cell communications in plants. Here, we investigated the function of the Populus CLV3/ESR-RELATED 47 (PttCLE47) gene during secondary growth and wood formation in hybrid aspen (Populus tremula × tremuloides) using an RNA interference (RNAi) approach. Expression of PttCLE47 peaks in the vascular cambium. Silencing of the PttCLE47 gene expression affected lateral expansion of stems and decreased apical height growth and leaf size. In particular, PttCLE47 RNAi trees exhibited a narrower secondary xylem zone with less xylem cells/cell file. The reduced radial growth phenotype also correlated with a reduced number of cambial cell layers. In agreement with these results, expression of several cambial regulator genes was downregulated in the stems of the transgenic trees in comparison with controls. Altogether, these results suggest that the PttCLE47 gene is a major positive regulator of cambial activity in hybrid aspen, mainly promoting the production of secondary xylem. Furthermore, in contrast to previously characterized CLE genes expressed in the wood-forming zone, PttCLE47 appears to be active at its site of expression.

Introduction

In plants, peptide‐receptor signalling modules have important roles in mediating cell‐to‐cell communications and interactions during growth, development, and responses to environmental stimuli. One of the best‐studied gene families encoding such small peptide ligands is the CLAVATA3 (CLV3)/EMBRYO SURROUNDING REGION (ESR)‐RELATED (CLE) gene family (Clark et al., 1995; Fletcher et al., 1999). CLE genes encode small proteins (c. 60–120 aa), which carry an N‐terminal signal peptide and a C‐terminal conserved CLE peptide domain of length 12–13 aa (Oelkers et al., 2008; Ohyama et al., 2008, 2009). The CLE peptide domain is excised from its prepropeptide by serine proteases and carboxypeptidases, and posttranslationally modified (i.e. hydroxylation and arabinosylation) to become biologically active (Fiers et al., 2005, 2006; Ito et al., 2006; Ni & Clark, 2006; Ohyama et al., 2008, 2009; Ni et al., 2011). In addition, the signal peptide is essential for the release of the active CLE peptide into the extracellular space, where it functions (Rojo et al., 2002).

In Arabidopsis thaliana, CLE genes are expressed with different tissue specificities and mainly interact with the plasma‐membrane‐associated LEUCINE‐RICH REPEAT RECEPTOR‐LIKE KINASEs (LRR‐RLKs) (Clark et al., 1995; Sharma et al., 2003; Hirakawa et al., 2008; Jun et al., 2010). Prominent examples from this peptide family include CLV3 and CLE40, which are involved in the regulation of stem cell pools in the shoot apical meristem and root apical meristem, respectively, in Arabidopsis (Clark et al., 1995; Brand et al., 2000; Schoof et al., 2000; Stahl & Simon, 2009; Stahl et al., 2009). Other than their roles in stem cell homoeostasis in apical meristems, CLE genes also regulate a variety of other biological processes, such as expansion of lateral roots in response to nitrogen availability (Araya et al., 2014), embryo and endosperm development (Fiume & Fletcher, 2012), phloem initiation (Ren et al., 2019) and protophloem development (Rodriguez‐Villalon et al., 2014), protoxylem development (Kondo et al., 2011; Qian et al., 2018), stomatal development (Qian et al., 2018) and closure upon dehydration stress (Takahashi et al., 2018; Zhang et al., 2018), pollen–pistil interactions (Endo et al., 2013), and autoregulation of nodulation (Okamoto et al., 2009, 2013; Mortier et al., 2010; Reid et al., 2011).

The secondary xylem (wood) and the secondary phloem of the plants are generated by the proliferative activity of the vascular cambium – a secondary meristematic tissue that contains the vascular stem cells. It has been shown that, similar to the apical meristems, the activity of the vascular stem cells in the procambium/cambium in Arabidopsis stems and hypocotyls is also mediated by an interaction between the CLE peptide TRACHEARY ELEMENT DIFFERENTIATION INHIBITORY FACTOR (TDIF)/CLE41/CLE44 and the LRR‐RLK PHLOEM INTERCALATED WITH XYLEM (PXY)/TDIF RECEPTOR (TDR) (Ito et al., 2006; Hirakawa et al., 2008; Etchells & Turner, 2010). CLE41 and CLE44 genes are expressed in the phloem tissues and neighbouring cells in the inflorescence stems and hypocotyl (Hirakawa et al., 2008; Etchells & Turner, 2010). The secreted TDIF/CLE41/CLE44 peptide is perceived by TDR/PXY, which shows an expression maximum in procambium and in the xylem side of the cambium (i.e. in the stem cell organizer) (Hirakawa et al., 2008; Etchells & Turner, 2010; Smetana et al., 2019). This interaction promotes the vascular cell divisions, suppresses the xylem cell specification, and controls the patterning of the vascular tissues (Ito et al., 2006; Hirakawa et al., 2008, 2010; Etchells & Turner, 2010; Etchells et al., 2012, 2013).

Reports suggested that the TDIF/CLE41/CLE44‐PXY/TDR signalling module is evolutionarily conserved in regulating the secondary growth and wood formation of Populus trees (Etchells et al., 2015; Kucukoglu et al., 2017). Aside from TDIF/CLE41/CLE44‐like genes, a recent study showed that the xylem‐produced peptide PtrCLE20 negatively regulates cambial activity in Populus (Zhu et al., 2019). Poplars are deciduous, hardwood trees and widely cultivated world‐wide as bioenergy feed stocks and for production of pulp, paper, packing, and woody material. Accompanied by the available sequenced genome (Tuskan et al., 2006), transformation and in vitro propagation techniques (Nilsson et al., 1992), as well as transcriptomic resources (Schrader et al., 2004; Sundell et al., 2017), it is one of the best model plants in understanding the molecular basis of tree growth and development, particularly the formation of wood. Other than the studies already mentioned, functional information regarding the roles of different CLE peptides in this important tree species is scarce. This prompted us to investigate the function of the cambium‐expressed CLE gene PttCLE47in hybrid aspen. Our results demonstrate that PttCLE47 has a role in controlling stem secondary growth by positively regulating the cell division activity of the vascular cambium in trees at its site of expression. It also affects overall growth and leaf size.

Materials and Methods

Plant material and growth conditions

Transgenic and wild‐type (WT) hybrid aspen (Populus tremula × tremuloides Michx, clone T89) plants were propagated from cuttings in tissue culture for 4 wk and transferred to soil. For phenotypic characterization they were grown under long day conditions (18 h : 6 h, light : dark, 22°C : 18°C) with 60–70% humidity for 3 months. During this period, they were rotated weekly to randomize environmental effects. Plant growth was assessed by measuring plant height, number of internodes, and stem width once a week for 6 wk. Height of the trees was measured from the shoot tips to the bottom of the stems. Stem width was measured and followed in the first visible internodes at the bottom of trees, 20 cm above the soil level. The number of internodes and leaves was counted omitting 3 cm from the shoot tips and a 20 cm long section above the soil (this part of the stem normally grows inside the tissue culture media during clonal propagation of trees, and upon transferring the plants to soil it grows slightly atypical with smaller, nonexpanding leaves independent of genotype; hence, it is omitted). In order to calculate the leaf areas, leaves (numbers 7 and 21, counted 3 cm below the shoot tip) were scanned and leaf area was calculated using imagej software (http://imagej.net). All statistical analyses were performed by GraphPad Prism v.8, using recommended tests. Phenotypes were compared using either an unpaired parametric t‐test (to compare means of two groups at a time; i.e. WT vs RNA interference (RNAi)‐n in 1 wk) by assuming a Gaussian distribution with Welch's correction (for unequal variance) or a Holm–Sidak multiple comparisons t‐test (to compare groups of means within each of the rows of data at a time; i.e. WT vs RNAi‐n over multiple weeks of growth), without assuming consistent standard deviations.

Cloning and transformation of hybrid aspen

The gene accession no. for PtCLE47 gene is Potri.017G074600.1 (Han et al., 2016). It was also annotated as PtrCLE25C (Zhu et al., 2019). Peptide sequences for PtCLE47 and AtCLE25 (AT3G28455) were obtained from Zhu et al. (2019). For RNAi plants, a 155 bp long RNAi fragment (also targeting the putative predicted active PttCLE47 peptide domain) was amplified from hybrid aspen stem complementary DNA (cDNA) using Phusion High‐Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA). The RNAi fragment was cloned via the Gateway Technology (Thermo Fisher Scientific) first into pDONR201 and further transferred into pK7GWIWG2 (I) (Karimi et al., 2002), using the Escherichia coli bacterial strain DH5α. The binary vector was transformed into Agrobacterium tumefaciens C58 strain GV3101 pMP90RK, and hybrid aspen trees were transformed as described previously (Nilsson et al., 1992). The primers utilized are listed in Supporting Information Table S1.

Gene expression analysis

For in silico gene expression analyses, expression data of PtCLE47, PtCLE24 (Potri.008G191500.1) and PtCLE28 (Potri.010G039800.1) were extracted from the ASPWOOD database (http://aspwood.popgenie.org; Sundell et al., 2017). For transcriptional analyses of PttCLE47 in different tissues of hybrid aspen, a cDNA sample set from 6‐month‐old trees described previously by Kucukoglu et al. (2017) was used. For subsequent gene expression analyses, stem segments from 18‐d‐old internodes of transgenic PttCLE47 RNAi trees and corresponding 18 d‐old WT internodes were harvested. As a control, stem segments from 11‐d‐old internodes of WT trees, which displayed similar thickness to 18 d‐old internodes of PttCLE47 RNAi trees, were also collected. For expression analysis in stocks segments of grafted plants, internodes 5 cm below the graft junctions were sampled. Total RNA was extracted using the cetyl trimethylammonium bromide method (Chang et al., 1993). Following the extraction, total RNA was treated with DNAse using the Ambion DNA‐free™ DNA Removal Kit (Thermo Fisher Scientific) and cDNA was synthesized using the iScript™ cDNA Synthesis Kit (Bio‐Rad) according to the manufacturer's instructions. Quantitative PCR (qPCR) analysis was performed in 10 µl reactions using the LightCycler® 480 SYBR Green I Master (Roche Life Science) in LightCycler® 480 instrument II (Roche Life Science). Each sample was analysed by three technical replicates. Detected expression levels were normalized against UBQ or the geometric mean of expression for four reference genes (18S, UBQ, 50S, TIP41‐like; Livak & Schmittgen, 2001). All statistical analyses were performed using GraphPad Prism v.8 for Mac OS X (GraphPad Inc.). Expression data were compared using an unpaired parametric t‐test (to compare means of two groups at a time, i.e. WT vs RNAi‐n) by assuming a Gaussian distribution with Welch's correction (for unequal variance). The primers utilized are listed in Table S1.

Histology

For histological analysis, stem segments from 18‐d‐old internodes of transgenic PttCLE47 RNAi trees and corresponding 18‐d‐old WT internodes were harvested. Stem segments from 11‐d‐old internodes of WT trees, which displayed similar thickness to 18‐d‐old internodes of PttCLE47 RNAi trees, were used as a control. These samples were fixed in a 5% formaldehyde, 5% acetic acid, and 50% ethanol mixture, gradually dehydrated, and embedded in Leica HistoResin (Leica Biosystems, Wetzlar, Germany). Thin cross‐sections (5 μm) were cut using a Leica RM 2055 rotary microtome (Leica Biosystems) and stained with toluidine blue and ruthenium red. Thick cross‐sections (100 μm) were cut using a vibrating‐blade microtome (Microm Microtech France, Brignais, France) and stained with safranin and Alcian blue, Bright‐field images were taken using a Leica DM2500 optical microscope (Leica Microsystems, Wetzlar, Germany). Secondary xylem and secondary phloem widths were measured using the imagej software (http://imagej.net), from the border of the cambium to the outermost cells of the pith and phloem fibres, respectively. Bark size was measured from the border of the cambium to the outermost cells of the cortex. The number of undifferentiated cambial cells and secondary xylem cells was calculated by counting 10 cell files and three cell files, respectively, from three trees per line. All statistical analyses were performed using GraphPad Prism v.8. Tissue thicknesses were compared using a Holm–Sidak multiple comparisons t‐test (to compare groups of means within each rows of data at a time, i.e. WT‐18D vs RNAi‐n‐18D for multiple characteristics, including thickness of secondary xylem, secondary phloem, and bark), without assuming consistent standard deviations. The number of cambial cells was compared using an unpaired parametric t‐test (to compare means of two groups at a time, i.e. WT vs RNAi‐n) by assuming a Gaussian distribution with Welch's correction (for unequal variance).

Grafting

Grafting experiments were performed as previously described (Nieminen et al., 2008). Briefly, transgenic and WT hybrid aspens were propagated from cuttings in tissue culture for 4 wk and transferred to soil. After 50 d, 5 cm long scions were excised and trimmed into a triangular shape. Simultaneously, longitudinal splits were made in stock stems, and trimmed scions were inserted. The scion–stock graft junction was secured with parafilm, and grafted plants were covered with plastic bags until after the graft junction to prevent drying. Bags were removed after 2 wk, when grafted scions started producing new leaves. Plant growth was assessed by measuring stem girths 5 cm below the graft junctions for 6 wk. Statistical analysis was performed using GraphPad prism v.8. Stem widths of stocks were compared using an unpaired parametric t‐test (to compare means of two groups at a time, i.e. WT/WT vs RNAi/RNAi) by assuming a Gaussian distribution with Welch's correction (for unequal variance).

Results

PttCLE47 is highly expressed in the cambium of hybrid aspen

The Populus trichocarpa genome encodes 52 CLE genes (Zhu et al., 2019), hereafter referred to as PtCLEs (for P. trichocarpa) or PttCLEs (for P. tremula × tremuloides, hybrid aspen). Recent studies indicated that 33 PtCLEs are expressed, with various levels of specificity, in the Populus stem, including secondary phloem, vascular cambium, expanding xylem, and maturing lignified xylem tissues (Han et al., 2016; Zhu et al., 2019). This suggests that members of the PtCLE family may potentially regulate a diversity of functions during vascular development in Populus trees, such as cambial activity, specification of xylem and phloem cell types, or secondary cell wall formation. Based on these gene expression profiles (Han et al., 2016; Zhu et al., 2019), we decided to focus on the previously uncharacterized PtCLE47 gene, which shows a strong expression gradient over the vascular cambium (Fig. 1a). We verified the transcript levels of PttCLE47 via qPCR analysis in different tissues of hybrid aspen trees, including shoot apices, leaves, stems, phloem‐cambium and xylem tissues, as well as roots (Fig. 1b). PttCLE47 transcript levels were most abundant in the phloem‐cambium samples, which is in agreement with the RNA‐sequencing data (Fig. 1a), suggesting that PttCLE47 may be involved in the regulation of vascular development in hybrid aspen. A lower but detectable PttCLE47 expression was also observed in shoot apices, young leaves, young stems, and roots (Fig. 1b).

Figure 1

PttCLE47 is highly expressed in the cambial zone of Populus. (a) Smoothed and micrometre‐scaled expression of PtCLE47 (black), PtCLE24 (turquoise), and PtCLE28 (brown) as retrieved from the AspWood RNA‐sequencing database (http://aspwood.popgenie.org). The y‐axis shows average variance‐stabilizing transformation (VST) expression, and the x‐axis shows tangential samples over the wood‐forming zone (T1‐nn). Secondary phloem, cambium, expanding xylem, and lignified xylem are marked. (b) Relative transcript abundance of PttCLE47. Expression levels were quantified via quantitative PCR in different tissues of hybrid aspen and normalized against UBQ reference gene expression. Average expression of PttCLE47 in all samples is equalized to one. Values are means ± SE (n = 3). (c) Sequence alignment of AtCLE25 and PtCLE47 peptides. Asterisks show conserved amino acid sequences. At, Arabidopsis thaliana; Pt, Populus trichocarpa.

Previous phylogenetic analyses (Han et al., 2016; Zhu et al., 2019) indicated that the PtCLE47 gene encodes a small A‐type peptide (subtype III) similar to Arabidopsis CLE25 (Fig. 1c). A‐type peptides include CLV3 and similar peptides based on sequence and functional similarity, whereas B‐type peptides contain only TDIF/CLE41/CLE44‐like peptides (Whitford et al., 2008). CLE25 is expressed in the vascular tissues of the shoot, root, and young leaf primordia in Arabidopsis (Jun et al., 2010; Takahashi et al., 2018), consistent with our findings. Interestingly, though, when investigated in detail, CLE25 expression was found to follow the development of the phloem cell lineage during vascular development (Ren et al., 2019). Moreover, knockout mutants of cle25 cause delayed protophloem differentiation in the primary roots of Arabidopsis, which suggests that CLE25 mainly regulates phloem formation. Though the expression of PtCLE47 peaks over the vascular cambium (Fig. 1a), its closest homologues PtCLE24 and PtCLE28 (Han et al., 2016) display slightly shifted expression patterns towards the differentiating phloem in trees (Fig. 1a). Therefore, whereas the PtCLE24 and PtCLE28 genes may represent the functional orthologues of the CLE25 gene in Populus, the PtCLE47 gene may have acquired a different function during Populus tree evolution.

Silencing of PttCLE47 affects apical/lateral growth, and leaf size in hybrid aspen

To investigate the function of PttCLE47 in trees, we generated transgenic hybrid aspen trees with RNAi‐mediated gene‐silencing of PttCLE47 gene expression (Figs 2, S1). A total of 14 transgenic lines were generated, and four lines (lines 1, 4, 8, and 13) with reduced expression of PttCLE47 compared with WT trees were chosen for further analysis (Fig. 2a). qPCR analysis of PttCLE47 transcript levels in stems of these four lines displayed that the expression of PttCLE47 was reduced to 7–17% of the WT levels. In Populus, PtCLE28 and PtCLE24 genes display high sequence similarity to PtCLE47 (Han et al., 2016). Expressions of these two genes were affected only minimally in the PttCLE47 RNAi plants (Fig. S2a,b), suggesting that the PttCLE47 RNAi construct is specific and does not target similar genes.

Figure 2

PttCLE47 RNA interference (RNAi) transgenic hybrid aspen trees display reduced growth. (a) RNAi‐mediated downregulation levels of the PttCLE47 gene in transgenic hybrid aspen plants. Relative transcript abundance of PttCLE47 was quantified via quantitative PCR and normalized against geometric means of the expression of four reference genes (18S, UBQ, 50S, and Tip41‐like). Values are means ± SE (n = 3). (b, c) PttCLE47 RNAi trees are shorter in comparison with wild‐type (WT) plants: (b) WT (left) and PttCLE47 RNAi trees (right); (c) plant height for 10‐wk‐old plants. (d) PttCLE47 RNAi trees are thinner in comparison with WT plants. Stem width for 10‐wk‐old plants. (e) Internode numbers for 10‐wk‐old plants. Values are means ± SE (n = 7–10 for PttCLE47 RNAi plants, n = 6 for WT). Statistical significances are determined using an unpaired parametric t‐test with Welch's correction. t‐test significance for different means: **, P < 0.01; ***, P < 0.001.

The PttCLE47 RNAi lines and corresponding WT trees were propagated from in vitro cuttings and grown on soil under glasshouse conditions. After 10 wk of growth, transgenic trees displayed a small but significant reduction in height growth (Figs 2b,c, S1a,b), and had thinner trunks (Figs 2d, S1c) in comparison with WT trees. However, internode numbers of PttCLE47 RNAi plants remained unchanged (Figs 2e, S1d), suggesting that the difference in height growth could be related to the differences in internode lengths rather than the initiation rate of leaves at the shoot apex. Taken together, these results suggest that PttCLE47 promotes both apical and lateral growth in hybrid aspen.

We also observed a difference between the sizes of the PttCLE47 RNAi and WT leaves (Fig. 3a). The decrease in the leaf size of the PttCLE47 RNAi trees was apparent both in young expanding leaves (leaf number 7) and fully grown leaves (leaf number 21), resulting in c. 40–50% reductions in final leaf size (Fig. 3b). Although this aspect has not been investigated further, it is worth mentioning that the shape of the PttCLE47 RNAi leaves was also slightly different than WT leaves, where RNAi leaves displayed more lobes in comparison. In accordance with a role in controlling leaf growth and shape, PttCLE47 is also expressed in the young expanding leaves (Fig. 1b).

Figure 3

PttCLE47 RNA interference (RNAi) transgenic hybrid aspen trees display reduced leaf size. (a) Leaves from wild‐type (WT; left) and PttCLE47 RNAi trees (right). (b) Leaf areas from WT and transgenic plants. Values are means ± SE (n = 7–8 for PttCLE47 RNAi plants, n = 12 for WT plants). Statistical significances are determined using the Holm–Sidak multiple comparisons t‐test, without assuming consistent SD. t‐test significance for different means: **, P < 0.01; ***, P < 0.001.

Reduced lateral growth of PttCLE47 RNAi trees is due to local effects in the stem

Even though our phenotyping results suggested that PttCLE47 RNAi and WT plants display similar shoot apical meristem activities (i.e. initiation rate of leaves at the shoot apex each week were not changed and, as a result, WT and RNAi plants display similar numbers of internodes; Figs 2e, S1d), leaves of the transgenic trees were smaller than those of WT trees (Fig. 3). This may have an effect on the total photosynthetic area of RNAi trees and, in turn, reduce the amount of carbon available for growth and wood formation. Therefore, to exclude the effects of reduced leaf area on the radial growth of PttCLE47 RNAi trees, we performed a reciprocal grafting experiment (Figs 4a, S3a). For this purpose, one transgenic line (line 8) was reciprocally grafted to WT, and lateral thickening of stocks below the graft junctions was monitored. At 8 wk after grafting, as expected, stocks of WT/WT (scion/stock) were still significantly thicker than stocks of RNAi/RNAi (Fig. 4a). This is consistent with the thin‐stemmed phenotype of the ungrafted PttCLE47 RNAi plants (Figs 2d, S1c). WT/RNAi stocks also displayed compromised stem thickening in comparison with WT/WT stocks, suggesting that WT scions could only partially rescue the thin‐stemmed phenotype of RNAi stocks. Thus, our data showed that reduced PttCLE47 activity during secondary growth rather than reduced leaf size is the major cause of the thin‐stemmed phenotype of the mutant trees. Interestingly, RNAi/WT stocks were also thinner than WT/WT stocks after this growth period, which may be caused by the systemic spread of the RNAi signals across the graft junction from transgenic scions to nontransgenic stocks (Palauqui et al., 1997; Voinnet & Baulcombe, 1997; Han & Grierson, 2008). This hypothesis was supported by the fact that expression of PttCLE47 was downregulated in stocks of RNAi/WT plants compared with stocks of WT/WT plants (Fig. S3b).

Figure 4

PttCLE47 RNA interference (RNAi) transgenic hybrid aspen trees display reduced secondary xylem development. (a) Reciprocal grafts between wild‐type (WT) and PttCLE47 RNAi trees (line 8). Graph shows the stock width 8 wk after the graft junction was formed. The WT scion only partially complements the RNAi stocks. Values are means ± SE (n = 7–11). (b) Total widths of the secondary xylem, secondary phloem, and bark. Measurements were taken from the 18‐d‐old or 11‐d‐old internodes (from the shoot tips) Values are means ± SE (n = 3). (c) Number of xylem cells/cell file in 18‐d‐old internodes from WT and PttCLE47 RNAi plants. Values are means ± SE (n = 9, three biological replicates/line). (d) Anatomical comparison between WT and PttCLE47 RNAi plants. Cross‐sections were taken from the 18‐d‐old internodes of the PttCLE47 RNAi plants and corresponding WT plants (18‐d‐old stem segments displaying similar developmental age, 11‐d‐old stem segments displaying similar developmental stage/size). Double‐sided arrows (white) indicate secondary xylem regions. Bars, 500 μm. Statistical significances are determined using an unpaired parametric t‐test with Welch's correction for (a, c) and the Holm–Sidak multiple comparisons t‐test, without assuming consistent SD, for (b). t‐test significance for different means: **, P < 0.01; ***, P < 0.001. 18D, 18 d old; 11D, 11 d old; Xm, secondary xylem.

Downregulation of PttCLE47 causes reduced wood formation in hybrid aspen

Next, we investigated the effects of PttCLE47 downregulation on vascular development by histological comparison of cross‐sections from stems of PttCLE47 RNAi and WT trees (Fig. 4). Our anatomical observations (Fig. 4d) and measurements of the secondary vascular tissues (Fig. 4b) revealed that, in comparison with the WT stems of similar age (18‐d‐old internodes, denoted as WT‐18D), stems of the transgenic trees (18‐d‐old internodes, denoted as RNAi‐18D) exhibited a narrower secondary xylem zone (Fig. 4b,d) with less xylem cells per radial cell file (Fig. 4c). The secondary xylem size of the transgenic trees (18‐d‐old internodes, denoted as RNAi‐18D) was similar to that of WT trees with an earlier developmental stage (11‐d‐old internodes, denoted as WT‐11D; Fig. 4b,d). On the other hand, the size of the secondary phloem in PttCLE47 RNAi trees remained unaffected (Fig. 4b,d). Interestingly, the development of the cortex region in the transgenic trees was also impeded, which contributed to the thin‐stemmed phenotype (denoted as bark; Fig. 4b). These results suggest that PttCLE47 acts mainly as a promoter of wood formation in hybrid aspen.

Downregulation of PttCLE47 causes reduced cambial activity in hybrid aspen

As the number of new xylem and phloem cells is determined by the periclinal divisions of the cambial cells, we next studied the structure of the cambial zone in the stems of PttCLE47 RNAi and WT trees (Fig. 5). The cambial cells, localized between the differentiating xylem and phloem cells, can be observed as flat, thin‐walled files of cells. Our data demonstrated that, in the PttCLE47 RNAi trees, the vascular cambium consists of fewer cell layers than in WT trees of similar age (Fig. 5a,b). Altogether, these results indicate that PttCLE47 is a positive regulator of cambial activity in hybrid aspen, and mainly promotes the secondary xylem production.

Figure 5

Reduced secondary xylem size results from the reduced cambial activity in PttCLE47 RNA interference (RNAi) transgenic hybrid aspen trees. (a) Anatomical comparison between cambiums of wild‐type (WT) and PttCLE47 RNAi plants. Cross‐sections were taken from 18‐d‐old transgenic stems and corresponding WT plants (18‐d‐old stem segments displaying similar developmental age). Bars, 100 μm. Double‐sided arrows (white) indicate cambial region. (b) Number of cambium cells/cell file in 18‐d‐old internodes from WT and PttCLE47 RNAi plants. Values are means ± SE (n = 30, three biological replicates/line). (c) Cambial regulator genes are downregulated in PttCLE47 RNAi lines. Relative transcript abundance of PttHB4, PttHB7, PttWOX4a/b, and PttPXYa were quantified via quantitative PCR and normalized against geometric means of the expression of four reference genes (18S, UBQ, 50S, and Tip41‐like). Values are means ± SE (n = 3). Statistical significances are determined using an unpaired parametric t‐test with Welch's correction against WT‐18D. t‐test significance for different means: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Cm, vascular cambium; 18D, 18 d old; 11D, 11 d old.

To further characterize the molecular function of PttCLE47, we studied the expression of important cambial regulator genes in the stems of PttCLE47 RNAi and WT trees (Fig. 5c). The expression levels of the class III homeodomain‐leucine zipper (HD‐Zip III) transcription factor genes PttHB4 (Zhu et al., 2018) and PttHB7 (Zhu et al., 2013) was dramatically reduced in the stems of RNAi lines, both with respect to similar age and developmental stage of WT stems (Fig. 5c). Additionally, expression of Populus WUSCHEL‐RELATED HOMEOBOX4‐like genes (PttWOX4a/b; Kucukoglu et al., 2017), which are positive regulators of cambial activity and identity in Populus, and their upstream regulator PttPXYa (Etchells et al., 2015; Kucukoglu et al., 2017), were also reduced in the stems of transgenic plants (Fig. 5c). Thus, these results suggest that PttCLE47 could act upstream of these HD‐Zip IIIs and the PXY‐WOX4 module in hybrid aspen.

Discussion

Previous work on the roles of CLE peptides has been mainly limited to Arabidopsis, and only little is known about their functions in different plant species, including trees. This motivated us to investigate the role of PttCLE47, a putative peptide‐encoding gene, in hybrid aspen. We showed that PttCLE47 is predominantly expressed in the vascular tissues, particularly in the vascular cambium of trees. To analyse the function of the PttCLE47 gene, we generated knockdown trees with decreased expression of the PttCLE47 gene. Repression of this gene led to many defects in the mutant plants, including reduced apical growth and leaf size and impaired radial growth. Through grafting, we showed that reduction in the leaf size of the transgenic plants cannot explain the decrease in the lateral growth. A detailed anatomical examination revealed that RNAi plants had reduced secondary xylem formation because the vascular cambium of these plants contains fewer cell layers than that of WT. Altogether, our results demonstrated that PttCLE47 gene is a major positive regulator of cambial activity and secondary xylem formation in hybrid aspen. Notably, the action of PttCLE47, likely through the active PttCLE47 peptide, seems to be cell autonomous for the vascular cambium, functioning locally at the site of expression. Although most CLE peptides act noncell autonomously – that is, CLV3 (Clark et al., 1995; Brand et al., 2000; Schoof et al., 2000), CLE40 (Stahl & Simon, 2009; Stahl et al., 2009), and TDIF/CLE41/CLE44 (Ito et al., 2006; Hirakawa et al., 2008; Ohyama et al., 2008; Etchells & Turner, 2010) – it has also previously been observed that both the ligand CLE45 and the receptor BARELY ANY MERISTEM 3 are active in the same tissue in the primary roots to control protophloem development (Rodriguez‐Villalon et al., 2014). Therefore, a similar mechanism may be active here as well. Notably, in the PttCLE47 RNAi trees, size of the cortex region was also slightly reduced compared with that of WT trees (Fig. 4b), which was partly responsible for the thin‐stemmed phenotype. Therefore, it will be interesting in the future to study the expression of PttCLE47 in the cortex to determine if PttCLE47 peptide functions locally in this region as it does in the vascular cambium, or if it works noncell autonomously from the vascular cambium to affect the cortex development.

Previous phylogenetic analysis revealed that PtCLE47 together with its homologous genes PtCLE24 and PtCLE28 groups with Arabidopsis CLE25 gene (Han et al., 2016). CLE25 mainly regulates phloem formation in Arabidopsis and is expressed in the phloem cell lineage of stems and roots (Ren et al., 2019). Interestingly, our results showed that, in comparison with developing phloem‐expressed PtCLE24 and PtCLE28, expression of PtCLE47 is diverged and highest in the vascular cambium of trees. This points to the fact that true orthologues of CLE25 in poplar are PtCLE24 and PtCLE28 genes. Analysis of the Populus genome assembly provided evidence for two whole genome duplication events: one as a result of the salicoid event roughly 65 Ma, and an older eurosid duplication event, which coincided with the divergence of Populus and Arabidopsis lineages (Tuskan et al., 2006). These duplication events may have contributed to the expansion of the CLE25 gene clade in poplar, subsequent changes in gene expression patterns between PtCLE47 vs PtCLE24 and PtCLE28, and rewiring of the gene regulatory networks downstream of these genes. As a result, this may have caused the functional differentiation of PtCLE47 during evolution.

Previous studies using Populus as a model system for wood formation identified some important regulators of cambial development, including PtrHB4 (Zhu et al., 2018), PtrHB7 (Zhu et al., 2013), PttWOX4a/b (Kucukoglu et al., 2017), and PttPXYa (Etchells et al., 2015). Hence, we investigated the expressions of these key players in our mutant trees. PtrHB4 encodes an HD‐Zip III transcription factor similar to PHABULOSA and PHAVOLUTA in Arabidopsis (Zhu et al., 2018). In Arabidopsis, HD‐Zip IIIs have recently been shown to position and maintain cambial stem cells (Smetana et al., 2019). In poplar, HD‐Zip III transcription factors are required for cambium development (Robischon et al., 2011; Zhu et al., 2013, 2018). In particular, PtrHB4 is critical for the formation of a closed vascular cambium ring from fascicular and interfascicular cambium (Zhu et al., 2018). Additionally, PtrHB7, a homologue of AtHB8, controls a balanced differentiation between secondary xylem and secondary phloem tissues during lateral growth in Populus, such that PtrHB7‐silenced plants display a reduction in xylem but increase in phloem size (Zhu et al., 2013). Consistent with our mutant phenotypes, the expressions of both PttHB7 and PttHB4 are downregulated in the stems of RNAi trees, both in comparison with WT stems at the same developmental stage or at the same age. PttWOX4a/b and PttPXYa genes, which are both key regulators of cambial activity and identity in Populus (Etchells et al., 2015; Kucukoglu et al., 2017), also showed a modest but significant downregulation in the stems of mutant trees. Taken together, these results indicate that PttCLE47 might be part of a complex network of regulators of cambium activity and secondary growth.

Finally, it appears that there are many CLE peptides that affect cambial activity from different sides of the cambium, including PttCLE41 on the phloem side (Etchells et al., 2015; Kucukoglu et al., 2017), PttCLE20 on the xylem side (Zhu et al., 2019), and PttCLE47 in the cambial zone itself (Fig. 6). Previously, it was shown that, by overexpressing PttCLE41 and its receptor PttPXY in their original expression domains, one can significantly improve woody biomass production. Hence, our findings on the proliferative role of another CLE peptide in regulating cambial development and radial stem growth may contribute to the development of more efficient plant biomass production systems if/when used together with their signalling receptors. However, to achieve this, a more detailed characterization is required to elucidate the exact mechanism of how PttCLE47 controls cambial development in Populus.

Figure 6

Model for regulation of cambial activity through CLAVATA3 (CLV3)/EMBRYO SURROUNDING REGION (ESR)‐RELATED (CLE) peptides in Populus. Phloem produced PttCLE41, and cambial‐expressed PttCLE47 regulates cambial activity (cambial cell proliferation) positively in Populus. Xylem‐derived PttCLE20 inhibits cambial activity negatively in Populus. Secondary phloem is indicated in orange, vascular cambium is indicated in blue, and secondary xylem is indicated in green. Arrows show peptide movement and positive or negative function in regulating cambial cell proliferation.

Author contributions

ON and MK designed the research. MK and SC performed the experiments, collected and analysed the data, and together with ON interpreted the results. MK, SC, BZ, APM, YH and ON co‐wrote the manuscript.

Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Fig. S1 PttCLE47 RNAi transgenic hybrid aspen trees display reduced growth (Extended graphs).

Fig. S2 Relative expression of PttCLE24 and PttCLE28 in PttCLE47 RNAi transgenic hybrid aspen lines.

Fig. S3 Reciprocal grafts between WT and PttCLE47 RNAi transgenic hybrid aspen trees.

Table S1 Primers used in this study.

Click here for additional data file.