The BOP-type co-transcriptional regulator NODULE ROOT1 promotes stem secondary growth of the tropical Cannabaceae tree Parasponia andersonii.

Tree stems undergo a massive secondary growth in which secondary xylem and phloem tissues arise from the vascular cambium. Vascular cambium activity is driven by endogenous developmental signalling cues and environmental stimuli. Current knowledge regarding the genetic regulation of cambium activity and secondary growth is still far from complete. The tropical Cannabaceae tree Parasponia andersonii is a non-legume research model of nitrogen-fixing root nodulation. Parasponia andersonii can be transformed efficiently, making it amenable for CRISPR-Cas9-mediated reverse genetics. We considered whether P. andersonii also could be used as a complementary research system to investigate tree-related traits, including secondary growth. We established a developmental map of stem secondary growth in P. andersonii plantlets. Subsequently, we showed that the expression of the co-transcriptional regulator PanNODULE ROOT1 (PanNOOT1) is essential for controlling this process. PanNOOT1 is orthologous to Arabidopsis thaliana BLADE-ON-PETIOLE1 (AtBOP1) and AtBOP2, which are involved in the meristem-to-organ-boundary maintenance. Moreover, in species forming nitrogen-fixing root nodules, NOOT1 is known to function as a key nodule identity gene. Parasponia andersonii CRISPR-Cas9 loss-of-function Pannoot1 mutants are altered in the development of the xylem and phloem tissues without apparent disturbance of the cambium organization and size. Transcriptomic analysis showed that the expression of key secondary growth-related genes is significantly down-regulated in Pannoot1 mutants. This allows us to conclude that PanNOOT1 positively contributes to the regulation of stem secondary growth. Our work also demonstrates that P. andersonii can serve as a tree research system.

INTRODUCTION

Mitotic activity of stem cell populations in the shoot apical meristem (SAM) and the root apical meristem (RAM) are responsible for the plant primary growth and enable the plant to grow indefinitely. Plant primary growth is also supported by the presence of a vascular system allowing exchanges along the plant body. In angiosperm and gymnosperm, the primary vasculature is derived from SAM activity and is organized as isolated fascicular bundles interspaced with immature parenchyma surrounding the central pith (Evert and Eichhorn, 2006; Little et al., 2002; Mazur et al., 2014; Zhu et al., 2018). Primary fascicular bundles consist of vascular stem cells (procambium) from which primary xylem and phloem vascular tissues are generated inward and outward, respectively. To compete for light, plants need to increase their height and consequently strengthen their stems. Many plant species have evolved a developmental process called secondary growth, allowing them to grow radially. During evolution, such a strategy was very successful and, as an example, allowed trees to dominate forest ecosystems.

Secondary growth depends on the vascular cambium (hereafter cambium) activity, a cylindrical sheath of dividing cells within the stem. The cambium originates from fusion of the fascicular procambium and interfascicular cambium. The latter is derived from the transdifferentiation of the interfascicular parenchyma cells (Barnett, 1981; Fischer et al., 2019; Helariutta and Bhalerao, 2003; Johnsson and Fischer, 2016; Larson, 1994; Mellerowicz et al., 2001; Miyashima et al., 2013; Nieminen et al., 2015). The cambium produces secondary xylem (wood) inward and secondary phloem (bast) outward (Evert and Eichhorn, 2006).

The activity of cambium responds to environmental and mechanical stimuli and is controlled by phytohormones and signaling peptides. Most phytohormones appear to play promotive roles in the regulation of cambium and secondary growth (Brackmann and Greb, 2014; Campbell and Turner, 2017; Fischer et al., 2019; Miyashima et al., 2013; Ragni and Greb, 2018). Cambium stem cell proliferation is also controlled by the peptide‐receptor‐transcription factor module TRACHEARY ELEMENT DIFFERENTIATION INHIBITORY FACTOR (TDIF) — PHLOEM INTERCALATED WITH XYLEM TDIF RECEPTOR (PXY‐TDR) — WUSCHEL‐RELATED HOMEOBOX 4 (WOX4). In the cambium, the type‐B TDIF peptides CLE41 and CLE44 bind the receptor PXY‐TDR and induce the expression of the transcription factor WOX4. WOX4 inhibits xylem cell identity acquisition and allows the proper delimitation, orientation and promotion of cambium proliferation (Etchells et al., 2013, 2015; Etchells and Turner, 2010; Fisher and Turner, 2007; Hirakawa et al., 2010; Hirakawa et al., 2008; Immanen et al., 2016; Ito et al., 2006; Ji et al., 2010; Kucukoglu et al., 2017; Liu et al., 2016; Nilsson et al., 2008; Schrader et al., 2004; Suer et al., 2011; Tuominen et al., 1997; Whitford et al., 2008; Zhu et al., 2019). Besides PXY‐TDR, other receptors also contribute to cambium regulation. The SOMATIC EMBRYOGENESIS RECEPTOR KINASEs (SERKs) function as PXY‐TDR co‐receptors (Zhang et al., 2016), whereas the PXY‐TDR homologous receptors PXY‐LIKE1 (PXL1) and PXL2 function synergistically to regulate vascular tissue development in the stem (Etchells et al., 2013; Fisher and Turner, 2007). Furthermore, REDUCED IN LATERAL GROWTH1 (RUL1) and MORE LATERAL GROWTH1 (MOL1) control cambium homeostasis (Agusti et al., 2011; Gursanscky et al., 2016) and the receptor‐kinases ERECTA (ER) and ERECTA‐LIKE1 (ERL1) redundantly control secondary growth by preventing premature initiation of xylem fibre differentiation in the hypocotyl (Ikematsu et al., 2017). Taken together, these studies highlight the importance of hormonal and peptide signaling modules in the regulation of cambium formation and activity.

Current knowledge regarding cambium and secondary growth genetic regulation is under progress but is still far from complete. Most studies have been conducted using either Populus species, Populus hybrids or the herbaceous model Arabidopsis thaliana. We considered whether we could use Parasponia andersonii as a complementary research system to obtain novel insights into the genetic regulation of secondary growth. Parasponia andersonii is a fast growing tree, for which efficient transformation protocols and a sequenced and annotated genome are available (van Velzen et al., 2018; Wardhani et al., 2019; van Zeijl et al., 2018). The Parasponia lineage represents five perennial evergreen tropical tree species that are native to the Malay Archipelago (Becking, 1992). The Parasponia lineage is part of the larger Trema genus, which belongs to the Cannabis family (Cannabaceae; order Rosales) (van Velzen et al., 2018; Yang et al., 2013). Cannabaceae includes ten genera and approximately 180 species consisting mostly of trees and shrubs (Yang et al., 2013). Parasponia and Trema species are pioneer plants with similar growth characteristics that thrive well on poor soils. Especially, Parasponia can colonize harsh landscapes post‐volcano eruptions (Ishaq et al., 2020). Also, Parasponia species are the only non‐leguminous plants that have the capacity to establish root nodules with diazotrophic rhizobia, making it an important comparative research system for investigating the evolution of this trait (Behm et al., 2014; van Velzen et al., 2018). Parasponia can grow up to 6 m a year, similar to that reported for Trema orientalis, which is also known as Nalita (Jahan and Mun, 2003; Trinick, 1980). Nalita wood is diffuse and porous and is considered as an alternative source of fiber for pulp production in the paper industry (Jahan and Mun, 2003; Jahan et al., 2007, 2010).

Parasponia andersonii possesses significant advantages for applying reverse genetic studies, especially in the context of tree development. Its genome is relatively small (563 Mbp) and has a high level of homozygosity (Holmer et al., 2019; van Velzen et al., 2018). Parasponia andersonii did not experience a recent whole genome duplication, as is the case in Salicoid, and, consequently, cases of gene redundancy are rare compared to more traditional model trees such as Populus species or Populus hybrids (Brunner et al., 2000; Tuskan et al., 2006). Besides, P. andersonii benefits from in vitro propagation and rooting procedures, from an efficient transformation procedure taking 2–3 months (transformation rate > 50%) and from an effective CRISPR‐Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR‐associated protein 9) genome editing strategy allowing the generation of biallelic recessive mutants (Bu et al., 2020; Rutten et al., 2020; van Velzen et al., 2018; Wardhani et al., 2019; van Zeijl et al., 2018). Moreover, P. andersonii reaches its sexual maturity within 5 months and seeds remain viable for years (Becking, 1992). Its genetic assets coupled with its dedicated toolbox make P. andersonii particularly suitable for the investigation of tree development via reverse genetic approach, and the availability of viable seeds enables the investigation of early tree developmental phases.

The NODULE ROOT (NOOT) — BLADE‐ON‐PETIOLE (BOP) — COCHLEATA (COCH) — LIKE genes (NBCL, also known as BOP‐LIKE) are plant specific developmental regulators encoding BTB/POZ (BROAD‐COMPLEX, TRAMTRACK and BRICK‐A‐BRACK/POXVIRUS and ZINC FINGER) and ANKYRIN domain repeats proteins (Couzigou et al., 2012). NBCL genes encode co‐transcriptional regulators that can also act as E3 ubiquitin ligase adapters regulating protein homeostasis (Jun et al., 2010; Zhang et al., 2017). NBCLs are conserved in both dicots and monocots, and were shown to be involved in a myriad of developmental processes (Dong et al., 2017; Hepworth and Pautot, 2015; Jost et al., 2016; Khan et al., 2014; Magne et al., 2020; Tavakol et al., 2015; Toriba et al., 2019; Wang et al., 2016). NBCL proteins are involved in boundary formation. These specific domains are important for partitioning meristematic domains from lateral organs (Aida and Tasaka, 2006a, 2006b; Barton, 2010; Zadnikova and Simon, 2014). In line with this, we hypothesized that NBCL genes are also critical for stem secondary growth. However, such functioning has never been reported.

In the present study, our objectives were two‐fold. First, we aimed to determine whether P. andersonii could be used as a tree research system. Second, we aimed to determine whether NBCL genes contribute to stem secondary growth. To achieve these objectives, we investigated the function of the single NBCL gene of P. andersonii, namely PanNODULE ROOT1 (PanNOOT1), during stem secondary growth. To do so, a stem secondary growth developmental map was created and combined with the gene expression profiles of key secondary growth regulators. PanNOOT1 displayed a complementary expression profile compared to secondary growth regulators, suggesting an antagonistic function during cambium initiation. Subsequent knockout mutagenesis of PanNOOT1 demonstrated that PanNOOT1 is an essential component of the stem secondary growth regulation.

RESULTS

A developmental map of cambium formation in P. andersonii

The establishment of cambium marks the transition from a juvenile stem towards a mature and fully functional stem. Once fascicular and interfascicular cambia merge and form a complete cambium ring, stem secondary growth starts and both secondary xylem and phloem tissues accumulate. Because of a lack of information regarding the dynamics of cambium formation in P. andersonii, we investigated this developmental process using the P. andersonii wild‐type reference genotype WU1‐14 (PanWU1‐14) (Op den Camp et al., 2011; van Velzen et al., 2018). To determine the dynamics of cambium formation in P. andersonii, an epicotyl developmental map was created using in vitro grown seedlings. Transversal sections of the first internode above the cotyledons (epicotyl) were histologically analyzed from plantlets after 13, 15, 18, 22, 26, 30, 35 and 42 days post‐germination (Figure 1). From 13 to 22 days, P. andersonii epicotyl vascular elements were organized as distinct poles. At these stages, vascular poles consisted of xylem and phloem tissues developing inward and outward from the fascicular cambium, respectively (Figure 1, panels 1–16). At 22 days, periclinal divisions initiated in the interfascicular parenchyma adjacent to the fascicular bundles. We noted that these cell divisions tended to occur synchronously in the different interfascicular parenchyma regions. These observations indicated that the interfascicular cambium started to establish (Figure 1, panels 13–16). At this stage, in vitro grown P. andersonii plantlets harboured four developed leaves and started to develop a fifth one (Figure S1). From day 26, the cambium was fully established, forming a complete ring of dividing cells. Once the cambium of P. andersonii epicotyl is established, xylogenesis occurs rapidly. Secondary xylem and phloem tissues accumulated inward and outward from the cambium, respectively (Figure 1, panels 17–32). Therefore, we concluded that, for in vitro grown P. andersonii seedlings, cambium formation starts at 22 days post‐germination, as a synchronized process. Ultimately, the structure of the wood in P. andersonii epicotyl is typical and consists of the regular angiosperm wood elements such as xylem fibres and xylem vessels regularly interspaced with xylem rays.

Figure 1

Developmental kinetics of P. andersonii vascular cambium establishment. Transversal sections of the first internode above cotyledons (epicotyl) of in vitro grown P. andersonii seedlings of genotype PanWU1‐14 at 13, 15, 18, 22, 26, 30, 35 and 42 days post‐germination. The first column shows entire epicotyl transversal sections. The second column shows magnifications focusing on vascular tissues organization. The third column provides simplified schemes of the vascular tissue organization generated from images in the second column . The fourth column shows magnifications focusing on the cambial zone. Panels 1–16, from 13 to 22 days, PanWU1‐14 epicotyl vascular elements are organized as isolated poles. Each vascular pole consists of xylem and phloem tissues developing inward and outward from the procambium (red dotted‐lines), respectively. Panels 13–16, At 22 days post‐germination, the first interfascicular cambium cell divisions occur (red asterisks). Panels 17–32, from 26 to 42 days, PanWU1‐14 epicotyl transversal sections display a complete ring of cambial cells (red dotted‐lines) with xylem and phloem accumulating inward and outward from the cambium respectively. p, phloem tissues; x, xylem tissues; red dotted‐lines, vascular cambium; red asterisks, interfascicular cambium cell divisions. Ratios in schemes of the right column indicate the number of transversal sections showing identical vascular tissues organization. Thickness = 7 µm. Scale bars = 100 µm

The expression of secondary growth developmental marker genes correlates with the dynamics of cambium formation

We considered whether the dynamics of P. andersonii secondary growth correlated with the transcriptional activation of cambium, xylem and phloem developmental marker genes. To identify such marker genes in P. andersonii, we used inferred orthogroup data available for P. andersonii, Populus trichocarpa, Eucalyptus grandis and A. thaliana and determined orthologous gene relation for known secondary growth marker genes (Emms and Kelly, 2015; van Velzen et al., 2018). In addition, we included other Populus species in our analysis. We identified 67 orthogroups that contained genes known to be involved in secondary growth, representing 90 genes in P. andersonii, 159 genes in P. trichocarpa, 140 genes in Populus tremula, 144 genes in Populus alba, 121 genes in E. grandis and 109 genes in A. thaliana (Table S1). Subsequently, the P. andersonii orthologs of known secondary growth marker genes were identified by additional and reciprocal blast (Basic Local Alignment Search Tool) analysis between P. andersonii, A. thaliana and P. trichocarpa, as well as by phylogenetic reconstruction analysis (Dataset S1). The P. andersonii genome was shown to contain direct orthologs for all secondary growth‐related genes tested here, except for one, XYLEM CYSTEINE PEPTIDASE2 (XCP2). Collectively, this suggests that the P. andersonii lineage is less prone for gene duplications compared to the Populus genus.

Following the identification of P. andersonii secondary growth‐related orthologous genes, gene expression profiling was performed for key secondary growth markers with epicotyl samples from 13, 15, 18, 22, 26, 30, 35 and 42 days post‐germination plantlets that were used to establish the cambium developmental map of P. andersonii. First, gene expression profiles were generated for the P. andersonii genes PanCLE41, PanPXY‐TDR, PanWOX4 and PanMOL1 regulating cambium activity (Figure 2a; Table S1; Dataset S1). Quantitative real‐time polymerase chain reaction (qRT‐PCR) analysis revealed that these four cambium activity marker genes were simultaneously induced in the epicotyl at 22 days post‐germination, concomitantly with the first interfascicular cambium cell divisions (Figure 1, panels 13–16; Figure 2a). PanCLE41, PanPXY‐TDR, PanWOX4 and PanMOL1 expression continued to increase until 26 days post‐germination and then remained stable (Figure 2a).

Figure 2

PanNOOT1 gene expression profile compared to vascular cambium, xylem and phloem marker gene expression profiles during P. andersonii stem secondary growth kinetics.

(a–d) qRT‐PCR gene expression profile of PanNOOT1 gene compared to vascular cambium, xylem and phloem developmental markers during the kinetic of P. andersonii stem secondary growth described in Figure 1. qRT‐PCR gene expression analysis was performed on the first internode above cotyledons (epicotyl) from in vitro grown P. andersonii plants at 13, 15, 18, 22, 26, 30, 35 and 42 days. Parasponia andersonii vascular cambium, xylem and phloem marker genes represent orthologs of vascular cambium, xylem and phloem marker genes described in A. thaliana and/or in Populus sp. (Table S1 and Dataset S1). (a) qRT‐PCR gene expression analysis of PanCLE41 (light blue curve), PanPXY‐TDR (blue curve) and PanWOX4 (dark blue curve) involved in the ligand‐receptor‐TF feedback loop systems regulating vascular cambium stem cell maintenance. qRT‐PCR gene expression analysis of PanMOL1 (red curve) involved in the regulation of interfascicular cambium cell proliferation. (b) qRT‐PCR gene expression analysis of the class I KNOX transcription factor genes PanSTM1 (light green curve), PanSTM2 (light green dotted curve) and PanBP (green curve) involved in vascular cambium stem cell maintenance. (c) qRT‐PCR gene expression analysis of PanSND1 (light blue curve), PanVND6 (light blue dotted curve), PanCNA (blue curve) and PanNST1 (blue dotted‐curve) involved in xylem development. qRT‐PCR gene expression analysis of PanAPL (red curve) involved in phloem development. d, qRT‐PCR gene expression analysis of PanNOOT1 (black curve). (a–d) Gene expression data were normalized against the constitutively expressed PanELONGATION FACTOR1α (PanEF1α) gene as well as against the expression levels from 13‐day‐old reference samples. (a, b and d) The y‐axis represents fold changes. (c) The y‐axis represents log2 (fold changes). Results represent thre mean ± sem from three biological replicates. Gene abbreviations: PanCLE41, PanCLAVATA3/ESR‐RELATED41 (PanWU01x14_078150); PanPXY‐TDR, PanPHLOEM INTERCALATED WITH XYLEM‐TDIF RECEPTOR (PanWU01x14_218900); PanWOX4, PanWUSCHEL RELATED HOMEOBOX4 (PanWU01x14_119590); PanMOL1, PanMORE LATERAL GROWTH1 (PanWU01x14_105020); PanSTM1, PanSHOOT MERISTEMLESS1 (PanWU01x14_211410); PanSTM2, PanSHOOT MERISTEMLESS2 (PanWU01x14_287890); PanBP, PanBREVIPEDICELLUS (PanWU01x14_033300); PanSND1, PanSECONDARY WALL‐ASSOCIATED NAC DOMAIN1 (PanWU01x14_056920); PanVND6, PanVASCULAR‐RELATED NAC‐DOMAIN6 (PanWU01x14_182640); PanAPL, PanALTERED PHLOEM DEVELOPMENT (PanWU01x14_155850); PanCNA, PanCORONA (PanWU01x14_195660); PanNST1, PanNAC SECONDARY WALL THICKENING PROMOTING FACTOR1 (PanWU01x14_041300); PanNOOT1, PanNODULE ROOT1 (PanWU01x14_292800)

In P. tremula and the hybrid P. tremula × P. alba, class I KNOTTED1‐LIKE HOMEOBOX (class I KNOX) genes promote cambium activity and regulate cambium daughter cell differentiation (Du et al., 2009; Groover et al., 2006; Schrader et al., 2004). We thus investigated the expression of the P. andersonii class I KNOX genes PanSTM1 and PanSTM2, orthologous to A. thaliana SHOOT MERISTEMLESS (AtSTM) and P. tremula × P. alba ARBORNOX1 (ARK1, Groover et al., 2006), as well as PanBP, orthologous to A. thaliana BREVIPEDICELLUS (AtBP) and P. tremula × P. alba ARBORNOX2 (ARK2, Du et al., 2009) (Figure 2b; Table S1; Dataset S1). PanSTM1, PanSTM2 and PanBP gene expression profiles were highly similar and induced in the epicotyl from 22 days post‐germination when interfascicular cambium starts to form (Figure 1, panels 13–16; Figure 2b). PanSTM1, PanSTM2 and PanBP expression continued to increase until 30 days post‐germination. At this stage, the cambium was fully developed (Figure 1, panels 21–24; Figure 2b). Then, from day 30 to 42, the expression of these three class I KNOX genes tended to decrease.

Together, these seven cambium marker genes showed a similar expression profile and were simultaneously up‐regulated in the epicotyl of 22‐day‐old plants concomitantly with the initiation of the cambium establishment. This suggests that, in P. andersonii, these genes might also represent regulators of cambium initiation and functioning. The finding that the expression of these genes in the epicotyl reaches a plateau at approxmately 26 days post‐germination indicates that, once the cambium is established, cambium activity reaches an equilibrium.

We next investigated the gene expression profiles of the key xylem and phloem developmental markers PanSND1, PanNST1, PanVND6, PanCNA and PanAPL, which are, respectively, orthologous to the A. thaliana SECONDARY WALL‐ASSOCIATED NAC‐DOMAIN PROTEIN 1 (AtSND1) and NAC SECONDARY WALL THICKENING PROMOTING FACTOR 1 (AtNST1) promoting xylem fibre differentiation and secondary cell wall thickening (Mitsuda et al., 2007; Zhong et al., 2006), to the A. thaliana VASCULAR‐RELATED NAC‐DOMAIN 6 (AtVND6), a master regulator of xylem differentiation (Kubo et al., 2005; Ohashi‐Ito et al., 2010; Yamaguchi et al., 2010), to the class III HD‐ZIP transcription factor CORONA (CNA) involved in xylem development and tracheary elements differentiation (Carlsbecker et al., 2010; Du et al., 2011; Ilegems et al., 2010; Prigge et al., 2005), and to the MYB coiled‐coil‐type transcription factor, ALTERED PHLOEM DEVELOPMENT (AtAPL), required for phloem identity acquisition, phloem differentiation and xylem identity inhibition in phloem (Bonke et al., 2003) (Figure 2c; Table S1; Dataset S1). qRT‐PCR analysis showed that these xylem and phloem developmental marker genes were also induced during P. andersonii secondary growth. The expression of these genes was sequentially up‐regulated and showed a maximum expression at approximately 30–35 days post‐germination, after which expression levels started to decrease. PanSND1 was among the earliest induced marker genes, detectable in the epicotyl at 15 days post‐germination. By contrast, PanVND6 and PanAPL were induced from 22 days post‐germination and, finally, PanCNA and PanNST1 were only induced from day 26 onwards. Although PanVND6, PanAPL and PanCNA showed similar maximum expression levels, between 2‐ to 4‐fold changes, PanSND1 and PanNST1 showed higher maximum expression levels, with between 16‐ and 64‐fold changes, respectively (Figure 2c).

In conclusion, all of the P. andersonii orthologs of cambium, xylem and phloem developmental marker genes identified in Populus species and/or A. thaliana were induced during P. andersonii secondary growth. The expression profiles of the different marker genes that we investigated were correlated with the developmental dynamics of P. andersonii secondary growth.

The NBCL gene PanNODULE ROOT1 is down‐regulated during cambium initiation but expressed in xylem, cambial and phloem tissues during secondary growth

Studies in A. thaliana have shown that the class I KNOX transcription factors AtSTM and AtBP repress the expression of the NBCL genes AtBOP1 and AtBOP2 to allow xylem fibre differentiation and vessel formation during root thickening (Liebsch et al., 2014; Woerlen et al., 2017). Available P. andersonii transcriptomic data showed that the AtBOP1‐AtBOP2 orthologous gene PanNOOT1 is expressed in several organs, including the stem (Figure S2) (van Velzen et al., 2018; www.parasponia.org). To confirm the presence of PanNOOT1 transcripts in P. andersonii stem and to determine the expression profile of PanNOOT1 during the establishment of the cambium, qRT‐PCR were performed on the RNA extracted from the epicotyl samples corresponding to the secondary growth kinetic studies (Figure 1). This revealed that PanNOOT1 displayed a characteristic expression profile in the developing epicotyl (Figure 2d). From 15 to 22 days, prior to the first interfascicular cambium cell divisions, the expression of PanNOOT1 fell drastically and decreased by 71% at 22 days. Then, from 22 days, the expression increased and reached a stable expression level from 30 days onwards, coinciding with a fully developed and functional cambium (Figure 2d). This indicates that the expression profiles of PanNOOT1 and cambium activity gene markers are complementary, suggesting antagonistic functions during cambium initiation (Figure 2). It is also suggested that molecular mechanisms involving the repression of PanNOOT1 by class I KNOX transcription factors might be conserved in tree species for the control of cambium initiation and functioning.

To determine the spatial expression pattern of PanNOOT1 in P. andersonii stem, we performed in situ hybridization (ISH) using the PanNOOT1 probe sets from Shen et al., 2020 and investigated transgenic P. andersonii lines expressing a PanNOOT1 promoter GUS reporter construct (promPanNOOT1:GUS:PanNOOT1ter). Because transgenic P. andersonii lines are maintained clonally (Wardhani et al., 2019) and because in vitro propagated plantlets grew less synchronized compared to seedlings, we investigared PanNOOT1 spatial expression in 7–10‐week‐old plants and analyzed the second and third fully elongated internodes, counted from the shoot apical meristem. Using both approaches, the expression of PanNOOT1 was detected in phloem parenchyma, in cambial zone, in differentiating secondary xylem cells, in differentiated secondary xylem fibres and in the older primary xylem, whereas it was not detected in cortex, in phloem sclerenchyma, in phloem sieve elements and in xylem vessels (Figure 3; Figure S3). ISH and promoter GUS reporter fusion approaches showed redundant spatial expression profiles, indicating that the 3.385 kb region upstream of PanNOOT1 may represent functional promoter elements. These results suggest that PanNOOT1 might play a role in the regulation of phloem, cambial and xylem tissues development.

Figure 3

PanNOOT1 in situ RNA hybridization and promoter GUS expression patterns in P. andersonii stem. The expression pattern of PanNOOT1 was determined by in situ hybridization and using transgenic lines expressing a PromPanNOOT1:GUS:NOOT1ter fusion.

(a,b) In situ hybridizations were performed on cross sections of the second fully elongated internode from the shoot apex of 10‐week‐old P. andersonii PanWU1‐14 plants. (a) Anti‐sense RNA probes targeting the nodule‐specific PanNF‐YA1 mRNAs (PanWU01x14_284830) served as negative control (Bu et al., 2020). No specific expression pattern or signal background were observed for PanNF‐YA1. (b) Specific PanNOOT1 anti‐sense RNA probes (red signals indicated by black arrowheads) were detected in phloem parenchyma, in cambial zone, in differentiating secondary xylem cells, in secondary xylem fibres and in the older primary xylem. Signals were not detected in cortex, in phloem sclerenchyma and sieve elements, nor in xylem vessels. (c) PanNOOT1 gene expression pattern in the third fully elongated internode from the shoot apex of 7‐week‐old P. andersonii PanWU1‐14 stable transformants expressing the GUS reporter fusion PromPanNOOT1:GUS:NOOT1ter. The X‐gluc staining (blue coloration indicated by white arrowheads) was detected in phloem parenchyma, in cambial zone, in differentiating secondary xylem cells, in secondary xylem fibres and in the older primary xylem. Signals were not detected in cortex, in phloem sclerenchyma and sieve elements, nor in xylem vessels. Black dotted lines indicate the frontier between the cambial zone and the phloem tissues. Co, cortex; Psc, phloem sclerenchyma; Pp, phloem parenchyma; Pse, phloem sieve elements; dX, differentiating xylem; Xf, xylem fibres; Xv, xylem vessels. Thickness: (a,b) 6 µm; (c) 7 µm. Scale bar: (a) 50 µm; (b,c) 25 µm

The loss‐of‐function of PanNOOT1 alters stem secondary growth

To determine whether PanNOOT1 has a critical role in P. andersonii stem secondary growth, CRISPR‐Cas9 loss‐of‐function mutants were generated. Using three guide RNAs that target the first exon of the PanNOOT1 gene, three independent mutant lines were identified (Pannoot1 A5, A10 and A29). All three Pannoot1 mutant lines presented a homozygous single nucleotide mutation at base pair 36 of the coding region, which is covered by guide RNA 1. Additionally, all three lines presented a second and wider deletion that varies in size, being either homozygous or biallelic (Figure S4). In all cases, this caused premature stop codons in the PanNOOT1 open reading frame. The Pannoot1 A5 and A10 alleles may encode a 26 amino acid truncated protein, whereas, in the case of the Pannoot1 A29 allele, the putatively encoded truncated protein has a length of 18 amino acids (Figure S4). This allowed us to consider Pannoot1 A5, A10 and A29 as knockout mutants.

To determine the consequences of the loss‐of‐function of PanNOOT1 on P. andersonii stem secondary growth, Pannoot1 A5, A10 and A29 internode diameters were compared with the wild‐type line PanWU1‐14 and the transgenic control line PanCtr‐44 (van Zeijl et al., 2018). However, the Pannoot1 mutants presented a very strong alteration of axillary shoot outgrowth (Figure S5; Shen et al., unpublished data). Such a strong developmental defect might impact the photosynthetic rate, as well as secondary growth, and therefore could generate a bias in our interpretation. To ascertain that the Pannoot1 secondary growth phenotype is a result of the loss‐of‐function of PanNOOT1 rather than to a lack of axillary shoots, we included additional controls consisting of PanWU1‐14 and PanCtr‐44 plants manually trimmed for their axillary shoots since their seedling stage (hereafter, trimmed PanWU1‐14 and trimmed PanCtr‐44). Plant genotypes were grown for 10 weeks under greenhouse conditions (28°C, 85% relative humidity), resulting in plants with approximately 10 internodes [numbered here from most basal internode (IN1) to upper internode (IN10)]. At this stage of development, we noted that basal internodes produced cork, an outermost stem‐protective tissue also called periderm (Figure S5), and we showed that P. andersonii developed a cork cambium also called phellogen (Figure S6). In upper internodes, from IN6 to IN10, trimmed PanWU1‐14 and PanCtr‐44 had significantly larger internode diameters compared to untrimmed PanWU1‐14 and PanCtr‐44 plants (Figure 4). This suggests that the absence of axillary shoots led to an increased secondary growth in the youngest internodes. However, this phenotype was not observed in Pannoot1 mutants, despite the absence of lateral branch outgrowth. By contrast, in the basal internodes, from IN1 to IN4, the diameter of internode of Pannoot1 A5, A10 and A29 mutant plants were significantly reduced compared to all trimmed and untrimmed control plants (Figure 4). The reduced secondary growth phenotype of Pannoot1 was also observed in older plants that were analyzed 20 weeks post‐planting. Of the five most basal internodes, the diameters of Pannoot1 A5 were significantly reduced relative to PanWU1‐14 (Figure S7). These results indicated that the loss‐of‐function of PanNOOT1 negatively affects P. andersonii stem secondary growth in an axillary shoot‐independent manner.

Figure 4

Parasponia andersonii Pannoot1 mutants present a reduced stem secondary growth. Measurement of stem diameters in Pannoot1 A5 (light blue bars), Pannoot1 A10 (blue bars) and Pannoot1 A29 (dark blue bars) compared to wild‐type PanWU1‐14 (white bars), trimmed wild‐type PanWU1‐14 (hatched white bars), transgenic control PanCtr‐44 (grey bars) and trimmed transgenic control PanCtr‐44 (hatched grey bars). Stem diameters were measured at the middle of internodes (IN) for the first 10 internodes, from the bottom to the top of 10‐week‐old plants. For PanWU1‐14, trimmed PanWU1‐14, PanCtr‐44, trimmed PanCtr‐44, Pannoot1 A5, Pannoot1 A10 and Pannoot1 A29, n = 15, 15, 15, 15, 30, 30 and 26 plants, respectively. Error bars represent the SD. Asterisks indicate significant differences relative to wild‐type PanWU1‐14 (*P < 1 × 10−2; **P < 1 × 10−3; ***P < 1 × 10−4; ****P < 1 × 10−5; *****P < 1 × 10−6; ******P < 1 × 10−7; Student’s t‐test)

To understand why the Pannoot1 mutant stems were thinner, hand sections of the most basal internodes from all genotypes were made and the whole stem, the xylem and the phloem tissues areas were measured. Whole stem, xylem and phloem areas were significantly reduced in the three Pannoot1 mutants compared to all of the control plants (Figure S8). These results indicated that the reduced secondary growth observed in Pannoot1 is caused by a reduction of both xylem and phloem tissues development. This might be either the result of a reduction of cell size and/or a reduction of the number of cell layers. To determine why xylem and phloem surfaces were reduced in Pannoot1 mutants, and to characterize the Pannoot1 stem tissues organization, thin resin sections were obtained from the most basal internodes of 10‐week‐old plants. Figure 5 shows representative sections indicating the reduced secondary growth in Pannoot1 A5, A10 and A29 compared to the different controls . Besides the apparent secondary growth reduction occurring in Pannoot1 mutants, the organization of the Pannoot1 stem tissues did not differ from control plants. All genotypes showed a wild‐type tissue organization consisting of a central pith, sequentially surrounded by the xylem, the cambial zone, the phloem, the cortex and the periderm (Figure 5). Based on resin sections, secondary growth developmental parameters were defined and assessed in detail. In Pannoot1 mutants, the size of the cambium cells, the number of cell layers present in the cambial zone and the size of the first thickened xylem fibre cells were not significantly different compared to the control plants (Figure 5h, i, j). However, the number of differentiated xylem cell layers (Figure 5k) and the size of the differentiated xylem vessels were significantly reduced in Pannoot1 mutants (Figure 5l). This suggests that the formation and the differentiation of the xylem cells is affected in Pannoot1 mutants. In addition, at the phloem side, histological analysis revealed a reduced number of phloem cell layers in Pannoot1 mutants compared to the control plants (Figure 5m).

Figure 5

Histological analysis of stem tissues reveals a reduced number of xylem and phloem cell layers in P. andersonii Pannoot1 mutants.

(a–g) Representative images of the most basal internode organizations of 10‐week‐old PanWU1‐14 (a), trimmed PanWU1‐14 (b), PanCtr‐44 (c), trimmed PanCtr‐44 (d), Pannoot1 A5 (e), Pannoot1 A10 (f) and Pannoot1 A29 (g) genotypes. Pd, periderm; Co, cortex; Ph, phloem; Cz, cambial zone; Xv, xylem vessel; Xy, xylem fibre; P, pith. Thickness = 5 µm. Scale bars = 500 µm.

(h–m) Detailed analysis of secondary growth developmental parameters in the most basal internode of 10‐week‐old Pannoot1 A5 (light blue bars), Pannoot1 A10 (blue bars) and Pannoot1 A29 (dark blue bars) relative to wild‐type PanWU1‐14 (white bars), trimmed wild‐type PanWU1‐14 (hatched white bars), transgenic control PanCtr‐44 (grey bars) and trimmed transgenic control PanCtr‐44 (hatched grey bars) plants. (h) Measurement of vascular cambium cells size. Anticlinally dividing stem cells were specifically measured.

(i) Quantification of the number of cell layers present in the cambial zone. The cells were measured along transects from anticlinally dividing vascular cambium cells until the first thickened xylem fibre cells.

(j) Measurement of the first thickened xylem fibre cells size.

(k) Quantification of the number of differentiated xylem cell layers. The cells were measured along transects from the first thickened xylem fibre cells until the pith.

(l) Measurement of xylem vessel cells size.

(m) Quantification of the number of cell layers present in the phloem. The number of phloemian cell layers were quantified along transects from anticlinally dividing stem cells until the first cortex cell layers. (h–m) The number of elements analyzed (n) is indicated on the right of each graph. Error bars represent SDs. Asterisks represent significant differences compared to PanWU1‐14 control plants (***P < 1 × 10−4, Student’s t‐test). For the number of differentiated xylem cell layers parameter (k), the statistical analysis was performed on all Pannoot1 data and compared to all control lines (P < 6 × 10−133, Student’s t‐test)

Despite cambial zone integrity and organization apparently not being affected in P. andersonii Pannoot1 mutant lines, we showed that the reduced secondary growth phenotype observed in these mutants is the result of a reduced number of cell layers in both xylem and phloem tissues. These results suggest that, in the stem of P. andersonii, PanNOOT1 promotes both xylem and phloem tissues development.

The loss‐of‐function of PanNOOT1 affects the expression of secondary growth‐related genes

Because AtBOP1 and AtBOP2 act as co‐transcriptional regulators in A. thaliana, we considered to what extent the loss‐of‐function of PanNOOT1 can affect gene expression in P. andersonii internodes. To identify and quantify the transcriptional role of PanNOOT1 in P. andersonii internodes, we focused on the first three fully elongated internodes (feIN1, feIN2 and feIN3) from the shoot apex of 8‐week‐old plants. Histological analysis revealed that, in feIN1, feIN2 and feIN3, the cambium was fully established and that internodes underwent secondary growth. In these young internodes, no significant phenotypic differences were observed between Pannoot1 mutant alleles and control plants (Figure S9). Because of their developmental resemblance, these Pannoot1 A5, A10 and PanCtr‐44 internode samples were chosen for RNA sequencing (RNA‐seq) approaches to reduce gene expression variation background associated with dissimilar internode developmental status and to better focus on the consequences associated with the loss‐of‐function of PanNOOT1. The corresponding transcriptomic data have been integrated into an interactive public website (https://parasponia.plantgenie.org).

Principal component analysis (PCA) of the P. andersonii internode transcriptomic data highlighted a robust difference between PanCtr‐44 and the two Pannoot1 A5 and A10 mutants, regardless of the internode position. Indeed, two clusters formed by PanCtr‐44 samples and Pannoot1 mutant samples diverge over the x‐axis (PC1, 29, 43%) (Figure 6a). The PCA also revealed a divergence between feIN1 versus feIN2 and feIN3 irrespective of the genotype, following the y‐axis (PC2, 17, 96%) (Figure 6a). These results suggest that both Pannoot1 mutants had similar gene expression profiles, which differed from the PanCtr‐44 control line. Furthermore, the internode age/position influenced gene expression in a genotype‐independent manner.

Figure 6

The P. andersonii Pannoot1 mutants are affected in the expression of secondary growth marker genes.

(a) Principal component analysis (PCA) plot of the transcriptomic data. The PCA analysis was performed on internode samples collected from 8‐week‐old PanCtr‐44 (red), Pannoot1 A5 (blue) and Pannoot1 A10 (green). From the shoot apex, the first internode that was not shorter than the next one was defined as the first fully elongated internode feIN1. Downward internodes were numbered feIN2 and feIN3, consecutively. The different elongated internode samples used in the analysis are indicated by distinct shapes: feIN1 (circle), feIN2 (triangle) and feIN3 (square). All samples consisted of three biological replicates. The PCA analysis was performed on 27 transcriptomes and over 37229 P. andersonii genes. The first two components are shown, representing 47% of the variation in all samples.

(b,c) Differentially expressed transcripts are grouped within four distinct gene expression patterns. (b) Tissue responsive genes grouped either in tissue responsive pattern I (437 DETs, P < 0.01) in which genes are significantly up‐regulated in older internodes compared to younger internodes, irrespective of the genotype, or in tissue responsive pattern II (391 DETs, P < 0.01) in which genes are significantly down‐regulated in older internodes compared to younger internodes, irrespective of the genotype. (c) Genotype responsive genes can group either in the genotype responsive pattern III (3501 DETs, P < 0.01) in which genes are significantly down‐regulated in Pannoot1 mutants compared to PanCtr‐44, irrespective of tissue, or in the genotype responsive pattern IV (3677 DETs, P < 0.01) in which genes are significantly up‐regulated in Pannoot1 mutants compared to PanCtr‐44, irrespective of tissue. TPM, transcripts per million; feIN, fully elongated internode; n, number of genes.

(d) A diagram showing the frequency of tissue and genotype responsive DETs among 90 P. andersonii genes putatively involved in secondary growth (Table S1). A detailed table showing DETs among those 90 genes is provided in Figure S14.

(e) Heatmap of significantly down‐regulated transcripts related to secondary growth regulation in Pannoot1 A5 and A10 compared to PanCtr‐44. Parasponia andersonii genes were named according to the literature for A. thaliana and P. andersonii gene accession numbers are given. Key down‐regulated secondary growth‐related genes were sub‐divided into six groups according to their main function described in the literature: potential PanNOOT1 downstream targets, boundary specification, vascular cambium activity, xylem differentiation, xylem secondary cell wall and phloem differentiation. The heatmap scale represents normalized TPM values. For each gene, the sample with the lowest TPM value was normalized as 0 and the sample with the highest TPM value was normalized as 100. Gene abbreviations: PanNOOT1, PanNODULE ROOT1; PanATH1, PanARABIDOPSIS THALIANA HOMEOBOX GENE1 (At4g32980); PanKNAT6, PanKNOTTED‐LIKE FROM ARABIDOPSIS THALIANA 6‐LIKE (At1g23380); PanLOF, PanLATERAL ORGAN FUSION (At1g69560, At1g26780); PanCLE41, PanCLAVATA3/ESR‐RELATED41; PanWOX4, PanWUSCHEL RELATED HOMEOBOX4; PanXIP1, PanXYLEM INTERMIXED WITH PHLOEM1; PanRUL1, PanREDUCED IN LATERAL GROWTH1; PanMOL1, PanMORE LATERAL GROWTH1; PanANT, PanAINTEGUMENTA; PanBP, PanBREVIPEDICELLUS; PanSTM2, PanSHOOT MERISTEMLESS2; PanPNY, PanPENNYWISE; PanPIN1, PanPINFORMED1; PanHK4, PanHISTIDINE KINASE4; PanCOI1, PanCORONATINE INSENSITIVE1; PanD27, PanDWARF27; PanCCD7, PanCAROTENOID CLEAVAGE DIOXYGENASE7; PanCCD8, PanCAROTENOID CLEAVAGE DIOXYGENASE8; PanGA20OX2, PanGIBBERELLIN 20 OXIDASE2; PanPXC1, PanPXY‐TDR‐CORRELATED1; PanPHB, PanPHABULOSA; PanSND1, PanSECONDARY WALL‐ASSOCIATED NAC‐DOMAIN1; PanMYB46, PanMYB DOMAIN PROTEIN46; PanMYB42, PanMYB DOMAIN PROTEIN42; PanAPL, PanALTERED PHLOEM DEVELOPMENT

The transcriptomic data revealed that the expression of PanNOOT1 progressively increased in feIN2 and feIN3 compared to feIN1; however, this increase was not statistically significant. The expression of PanNOOT1 was drastically reduced in both Pannoot1 mutant lines compared to the control, suggesting either that the aberrant PanNOOT1 transcripts are degraded or that PanNOOT1 exerts a positive feedback loop on its own transcription (Figure 6e; Figure S10).

Differentially expressed transcripts (DETs) were grouped into four distinct categories of gene expression patterns (see below). Also, because transcription factors are essential regulators of plant developmental processes, including secondary growth (Chao et al., 2019), we explored the number of differentially expressed transcription factors in these four categories: Cat. I, genes that were significantly up‐regulated in feIN2 and feIN3 compared to feIN1, irrespective of the genotypes (437 DETs, P < 0.01, among which there were putative 44 transcription factors); Cat. II, genes that were significantly down‐regulated in feIN2 and feIN3 compared to feIN1, irrespective of the genotypes (391 DETs, P < 0.01, among which there were putative 27 transcription factors); Cat. III, genes that were significantly down‐regulated in Pannoot1 mutants compared to PanCtr‐44, irrespective of the internode positions (3501 DETs, P < 0.01, among which there were putative 222 transcription factors); and Cat. IV, genes that were significantly up‐regulated in Pannoot1 mutants compared to PanCtr‐44, irrespective of the internode positions (3677 DETs, P < 0.01, among which there were putative 215 transcription factors). There were no DETs that varied over both genotypes and internode positions (Figure 6b and c; Figure S11; Figure S12; Dataset S2; Dataset S3). These data indicate that the loss‐of‐function of PanNOOT1 has a broad deleterious effect on transcriptional regulation in developing stem internodes.

In A. thaliana meristem maintenance and flowering, AtBOP1 has been shown to directly promote the expression of ARABIDOPSIS THALIANA HOMEOBOX GENE1 (AtATH1) and to indirectly promote the expression of the class I KNOX transcription factor gene KNOTTED‐LIKE FROM ARABIDOPSIS THALIANA6 (AtKNAT6; Khan et al., 2015). In line with this, our transcriptomic analysis revealed that the closest orthologs of AtATH1 and AtKNAT6 in P. andersonii, namely PanATH1 and PanKNAT6, were significantly down‐regulated in Pannoot1 mutants (Figure 6e). These results support previous findings obtained in A. thaliana and suggest that PanNOOT1 might also promote the expression of PanATH1 and PanKNAT6 in this context of stem secondary growth.

NBCL genes, together with other meristem‐to‐organ‐boundary genes, regulate the boundaries between SAM and lateral organs. It is possible that NBCL also plays a role in defining cambium boundaries, similarly to that in SAM. The A. thaliana LATERAL ORGAN FUSION (AtLOF) genes are involved in SAM boundary specification and regulation. Here, we found that the single ortholog of AtLOF genes in P. andersonii, named PanLOF, was significantly down‐regulated in the Pannoot1 mutants (Figure 6e). Together with the loss‐of‐function of PanNOOT1, the down‐regulation of PanLOF might reflect a mis‐regulation of the boundaries existing between the cambium and the adjacent xylem and phloem tissues.

Secondary cell walls of xylem fibre and vessel cells are mainly composed of lignin (20–30%), together with cellulose (40–50%) and hemicellulose (25–30%) (Pradhan Mitra and Loqué, 2014). In A. thaliana and Gossypium hirsutum, NBCL positively regulates the expression of lignin metabolism‐related genes and the deposition of lignin in stems (Khan et al., 2012; Zhang et al., 2019). To determine whether PanNOOT1 also contributes to this process in P. andersonii, we investigated the expression of key lignin biosynthesis‐related genes in the transcriptomic data and tested the deposition of lignin using phloroglucinol‐HCl staining. In the first three fully elongated internodes (feIN1, feIN2 and feIN3) from the shoot apex of 8‐week‐old plants, transcriptomic data showed that several P. andersonii putative homologs of the lignin metabolism‐related genes were significantly down‐regulated in Pannoot1 internodes. We found that the P. andersonii PHENYLALANINE AMMONIA‐LYASE1 (PanPAL1), CINNAMATE 4‐HYDROXYLASE (PanC4H), 4‐COUMARATE COA LIGASE1 (Pan4CL1), COUMARATE 3‐HYDROXYLASE (PanC3H1), CAFFEOYL COENZYME A O‐METHYLTRANSFERASE1 (PanCCOAOMT1) and CINNAMOYL COA REDUCTASE1 (PanCCR1) were significantly down‐regulated in Pannoot1 mutants (Figure S13f). These results indicate that, in Pannoot1 mutants, the lignin biosynthesis pathway is affected, even though no apparent difference in lignified cells could be observed in 16‐week‐old Pannoot1 mutant plants (Figure S13a–e).

Next, we investigated the 90 P. andersonii genes belonging to 67 orthogroups putatively related to stem secondary growth regulation (Table S1). First, we considered whether the expression of these genes is influenced by internode position. Independently of the genotype, we found that the majority of these genes were not significantly differentially expressed in feIN2 and feIN3 compared to feIN1. Only 1% (1/90 genes) and 4% (4/90 genes) of these genes were significantly up‐ or down‐regulated, respectively. This is consistent with the findings obtained for the hybrid Populus deltoides × Populus euramericana, which revealed only 183 DETs when comparing two successive internodes (Chao et al., 2019). Among DETs in P. andersonii internodes, we found that PanSTM1 was significantly up‐regulated in feIN2 and feIN3 compared to feIN1 and that PanTARGET OF MONOPTEROS5‐LIKE1 (PanT5L1), PanVND6, PanXYLEM CYSTEINE PEPTIDASE (PanXCP) and PanCONSTITUTIVE PHOTOMORPHOGENIC DWARF (PanCPD) were significantly down‐regulated in feIN2 and feIN3 compared to feIN1 (Figure 6d, Figure S14). This small fraction of tissue responsive‐DETs suggests that the internode samples (feIN1, feIN2 and feIN3) were relatively homogeneous in terms of developmental status. However, in contrast to the minor effect observed for the impact of internode position on gene expression, approximately half of the putative secondary growth‐related genes were significantly mis‐regulated in Pannoot1 mutants compared to PanCtr‐44. We observed that 30% (27/90 genes) and 20% (18/90 genes) of the genes were significantly down‐ or up‐regulated in Pannoot1, respectively (Figure 6d; Figure S14). Among the genes that were significantly down‐regulated in Pannoot1 compared to PanCtr‐44, several were involved in the control of cambium activity, such as PanCLE41, PanWOX4, PanXIP1 (PanXYLEM INTERMIXED WITH PHLOEM1), PanRUL1, PanMOL1, PanANT (PanAINTEGUMENTA), PanBP, PanSTM2 and PanPNY (PanPENNYWISE). This suggests that PanNOOT1 positively contributes to the regulation of cambium activity genes (Figure 6e). However, we found that the expression of PanSTM1 and PanPXY‐TDR was not significantly different in Pannoot1 relative to PanCtr‐44, suggesting that the transcriptional regulation of these important cambium regulators is PanNOOT1‐independent (Figure S14). In addition, we found down‐regulated genes related to phytohormones, which act positively on cambium activity, notably auxin, cytokinin, jasmonic acid and strigolactone (Figure 6e). These results are in accordance with the reduced secondary growth phenotype of Pannoot1 and also suggest that the cambium activity of Pannoot1 mutants is impaired despite the absence of any obvious phenotype in the cambial zone itself. Besides cambium activity‐related genes, other actors participating in xylem differentiation, xylem secondary cell wall deposition or phloem differentiation were found to be down‐regulated in Pannoot1 (Figure 6e). Taken together, the down‐regulation of all of these important developmental markers is consistent with the reduced secondary growth phenotype observed in Pannoot1 mutants. These results highlight the positive contribution of PanNOOT1 with respect to the promotion of cambium activity and xylogenesis.

DISCUSSION

Stem secondary growth is a key characteristic of trees. We investigated this trait in the tropical Cannabaceae tree species P. andersonii. Serial sectioning of the epicotyl showed the formation of the cambium at 22 days post‐germination. This developmental process is associated with an increased expression of secondary growth developmental marker genes and especially cambium activity marker genes. Interestingly, the NBCL gene PanNOOT1 showed a complementary expression profile relative to cambium activity marker genes suggesting antagonistic functions during cambium formation. By exploiting the transformation potential of P. andersonii, PanNOOT1:GUS reporter lines and CRISPR‐Cas9 mutants were generated. We found that the expression pattern of PanNOOT1 correlated with the secondary growth defect of the Pannoot1 knockout mutants. Indeed, PanNOOT1 was transcribed in the cambial zone, in phloem parenchyma, in differentiating secondary xylem cells and in differentiated secondary xylem fibres, suggesting that PanNOOT1 acts as a positive regulator of secondary growth. A similar expression pattern was also found in P. tremula and birch (Betula pendula) by high‐spatial‐resolution gene expression studies (Alonso‐Serra et al., 2019; Sundell et al., 2017). In both species, the NBCL genes are expressed in developing phloem and in vascular cambium, as well as in developing and mature xylem tissues (Figure S15). Furthermore, we found that PanNOOT1 is required for the correct expression of secondary growth‐ and lignin biosynthesis‐related genes. Taken together, these studies unveiled a novel function for a NBCL gene in stem secondary growth and demonstrated that P. andersonii can serve as a tree research system.

Populus species and hybrids are commonly used to explore the regulation of secondary growth in trees. The transformation efficiencies of Populus species are relatively low (2–16%) but are more efficient in hybrids (10–40%). Usually, transformed shoots are obtained within 1–3 months and rooted transgenic plantlets within 3–8 months depending on the genotype (Cseke et al., 2007; De Block, 1990; Han et al., 2000; Yevtushenko and Misra, 2010). Only recently has progress been made to transform the recalcitrant Populus model P. trichocarpa Nisqually‐1. Nisqually‐1 transformation efficiency now reaches 27%, transformed shoots are obtained in 1 month and obtaining rooted transgenic plantlets takes 2 months (Li et al., 2017). Despite this recent advance, many transgenesis experiments are still conducted in poplar hybrids. However, CRISPR‐Cas9 reverse genetics in such lines is complicated when considering the whole genome duplication that occurred in the salicoid clade (Brunner et al., 2000). This duplication resulted in a large fraction of paralogous gene pairs rendering the genomes of Populus species and especially hybrids as relatively complex. As an example, there are approximately 8000 duplicated gene pairs in P. trichocarpa (Tuskan et al., 2006). In many cases, paralogous genes may act redundantly, requiring the knockout or knockdown of multiple gene copies before mutant phenotypes can be observed (Bruegmann et al., 2019). Therefore, the functional characterization of genes in Populus was often limited to over‐expression analysis and/or comparative studies in A. thaliana (Jin et al., 2017; Lu et al., 2013; McCarthy et al., 2010; Zhong et al., 2010, 2013). By contrast, the P. andersonii genome organization is less complex, as we visualized by characterizing 67 orthogroups of putative secondary growth‐related genes. In P. andersonii, these represent 90 genes, whereas in, P. trichocarpa, these represent 159 genes. This paralogous gene problem is also apparent for NBCL genes. Populus species have two NBCL paralogs, called BLADE‐ON‐PETIOLE‐LIKE1 and 2 (BPL1 and BPL2; Magne et al., 2020), whereas P. andersonii only possesses a single gene, namely PanNOOT1. Besides the reduced genome complexity of P. andersonii, efficient protocols for in vitro propagation, transformation and CRISPR‐Cas9 genome editing are also available for this species (Wardhani et al., 2019; van Zeijl et al., 2018). Because P. andersonii is fast growing, self‐compatible and sets seeds within approximately 5 months that can be stored for many years, this tree species is amenable for genetic dissection of tree‐specific traits.

In P. andersonii, cambium formation and subsequent secondary growth occur within approximately 3 weeks post‐germination. Here, we made a time series of sections visualizing this process in the epicotyl of a tree seedling. The transition from primary to secondary growth has also been investigated in the hybrid P. deltoides × P. euramericana but, because working with Populus seeds is not possible, Chao et al. (2019) instead used successive internodes with different stages of development. Also, the herbaceous model plant A. thaliana can undergo secondary growth in root, hypocotyl and stem. Cambium formation and secondary growth were investigated using this species (Fischer et al., 2019; Helariutta and Bhalerao, 2003; Johnsson and Fischer, 2016; Miyashima et al., 2013; Nieminen et al., 2015). Time series of sections visualizing A. thaliana hypocotyl secondary growth were first performed by Chaffey et al., 2002. Subsequently, Sankar et al., 2014 provided a high‐resolution atlas for A. thaliana hypocotyl secondary growth using two different ecotypes. In A. thaliana hypocotyl, the cambium formed approximately 1 week after germination; however, the flowering‐induced acceleration of secondary growth only occurs after 1–2 months of growth depending on the genotype and the photoperiod conditions (Ikematsu et al., 2017; Ragni et al., 2011; Sibout et al., 2008). Finally, A. thaliana secondary growth will remain limited compared to trees. This suggests that P. andersonii represents a relevant model for investigating cambium formation and subsequent secondary growth.

The P. andersonii Pannoot1 mutants have smaller xylem vessels, whereas the fibre cells are not affected. Recent studies in a P. trichocarpa × P. deltoids mapping population identified the potassium channel encoding gene ENLARGED VESSEL ELEMENT (EVE) as a regulator of xylem vessels dimension (Ribeiro et al., 2020). EVE expression is positively controlled by SND1 and we noted that the P. andersonii PanSND1 gene (PanWU1x14_056920) was down‐regulated in Pannoot1 mutant internodes (Figure 6e). We considered whether the down‐regulation of PanSND1 in Pannoot1 mutants might affect the expression of the P. andersonii EVE orthologous gene, which may explain the smaller xylem vessel phenotype. However, the closest ortholog of P. trichocarpa EVE in P. andersonii (PanWU01x14_222300) was up‐regulated in the internodes of Pannoot1 mutants. This suggests that PanNOOT1 may function in parallel or downstream of an EVE‐controlled pathway.

Research in different plant species has shown that one of the major roles of the NBCL proteins is to repress meristematic activity and promote adjacent tissues initiation and differentiation (Hepworth and Pautot, 2015; Wang et al., 2016; Zadnikova and Simon, 2014). NBCL proteins are involved in the differentiation and the patterning of several organs, such as stipules, nectaries, ligule, leaves, internodes, floral meristems, flowers, abscission zones, hypocotyls, roots or nodules (Couzigou et al., 2016; Couzigou et al., 2012; Ha et al., 2003, 2004, 2007; Hepworth et al., 2005; Khan et al., 2015; Khan et al., 2012; Liebsch et al., 2014; Magne et al., 2018; McKim et al., 2008; Norberg et al., 2005; Tavakol et al., 2015; Toriba et al., 2019; Woerlen et al., 2017; Yaxley et al., 2001). Consistent with this role in promoting organ differentiation and patterning, in the present study, we found that PanNOOT1 promotes both xylem and phloem development in stem, and also that PanNOOT1 is required for the correct expression of key secondary growth‐ and lignin biosynthesis‐related genes. These results are consistent with recent studies reporting that the differentiation of xylem cells is delayed in the primary root of Medicago truncatula Mtnoot1 mutants (Shen et al., 2019). However, studies in A. thaliana hypocotyls and roots showed that AtBOP1 and AtBOP2 repress xylem differentiation (Woerlen et al., 2017). This divergence in phenotype between A. thaliana roots, M. truncatula roots and P. andersonii stems implies that xylem formation is possibly not a direct readout of NBCL functioning but, instead, comprises an indirect effect and depends on the (different) interacting targets. Such interacting target proteins may vary depending tissues and/or plant species. Promoter GUS reporter studies in A. thaliana roots and P. andersonii stems revealed a different spatiotemporal regulation between AtBOP1/AtBOP2 and PanNOOT1 in vascular tissues. In P. andersonii stems, PanNOOT1 expression was detected in phloem parenchyma, in the cambial zone and in differentiating secondary xylem tissues, whereas, in A. thaliana roots, AtBOP1 and AtBOP2 were expressed in secondary phloem but not detected in the vascular cambium, nor in secondary xylem (Woerlen et al., 2017). Such divergence in transcriptional regulation may contribute indirectly to the difference in phenotype in secondary xylem formation.

Studies in A. thaliana showed that AtBOP1/AtBOP2 expression is repressed by homeodomain transcriptional regulators from the THREE‐AMINO‐ACID‐LOOP‐EXTENSION (TALE) family, such as the class I KNOX: BP and STM, and the BEL1‐like: PENNYWISE and POUND‐FOOLISH during meristem maintenance, flowering, and secondary xylem formation in roots (Khan et al., 2015; Woerlen et al., 2017). For example, expression of AtBP in root xylem represses AtBOP1/2 expression in this tissue (Woerlen et al., 2017). In agreement with such a repressing role of KNOX homeodomain transcriptional regulators, the NBCL gene BPL1 is significantly down‐regulated in the stem tissues of P. tremula × P. alba hybrid lines over‐expressing ARK1 (orthologous to STM) (Liu et al., 2015). Nevertheless, chromatin immunoprecipitation‐sequencing analysis using ARK1 as bait did not reveal that BPL1 or BPL2 are direct transcriptional targets (Liu et al., 2015). This suggests that, in the stem of the P. tremula × P. alba hybrid, ARK1 might repress the expression of BPL1 by an indirect mechanism. In addition, in P. tremula x P. alba, over‐expression of ARK1 and ARK2 (orthologous to BP) increases the size of the cambial zone and alters secondary growth. In these over‐expressing mutants, the differentiation of secondary xylem tracheary elements, fibre cells and secondary phloem is inhibited. The over‐expression of ARK1 and ARK2 is also associated with the down‐regulation of secondary growth‐related genes, including cell differentiation and hormonal regulation. Consistently, ark2 mutants obtained by synthetic miRNA‐suppression showed an increased secondary growth (Du et al., 2009; Groover et al., 2006). The phenotypes of the ARK over‐expressor mutants are reminiscent of those found for the P. andersonii Pannoot1 mutants because they both show an alteration of secondary growth. Taken together, these results suggest that, in tree stem secondary growth, STM/ARK1 and BP/ARK2 might act as NBCL repressors and that NBCL genes promote tree secondary growth as previously hypothesized in Khan et al., 2012. Based on the results obtained in P. andersonii and on the data available for A. thaliana, P. tremula × P. alba and G. hirsutum, we have proposed a regulatory model for PanNOOT1 in tree secondary growth (Figure S16). In this model, PanSTM1, PanSTM2 and PanBP act as indirect repressors of PanNOOT1. PanNOOT1, potentially through the activation of PanKNAT6 and PanATH1 expression, promotes stem secondary growth and the activation of lignin metabolism‐related genes.

In the present study, we have investigated to what extent P. andersonii could represent an alternative tree research model for untangling the genetic regulation underlying stem secondary growth. We have demonstrated the feasibility of investigating stem secondary growth using P. andersonii and show that this tree species represents a suitable model for applying reverse genetics. As a proof‐of‐concept, we knocked out and investigated the single NBCL gene of P. andersonii, PanNOOT1, and made a nbcl mutant in a tree species. This unveiled a novel promotive function for a NBCL gene in stem secondary growth.

EXPERIMENTAL PROCEDURES

Parasponia andersonii in vitro growing conditions from seeds

Mature brownish berries were collected from P. andersonii PanWU1‐14 trees and then the seeds were ridded of their skins by scratching. Seeds were disinfected using commercial sodium hypochlorite supplemented with a droplet of soap for 20 min and washed six times in sterile water. Seed germination was induced by continuous thermic cycles at 28°C for 4 h followed by 7°C for 4 h over 10 days. Seeds were sown on Schenk and Hildebrandt (SH) medium (pH 5.8), supplemented with 8 g L−1 Daishin agar. Plants were grown in a growth cabinet (Elbanton, Kerkdriel, The Netherlands) at 28°C under a 16:8 h light/dark photocycle.

Parasponia andersonii growing conditions for secondary growth assays

After in vitro propagation on propagation medium (SH salts, 3.2 g L−1; SH vitamins, 1 g L−1; sucrose, 20 g L−1; BAP, 1 µg ml−1; IBA, 0.1 µg ml−1; MES, 3 mm, pH 5.8; Daishin agar, 8 g L−1), P. andersonii shoots were rooted in vitro for 1 month on rooting medium (SH salts, 3.2 g L−1; SH vitamins, 1 g L−1; sucrose, 10 g L−1; IBA, 1 µg ml−1; NAA, 0.1 µg ml−1; MES, 3 mm, pH 5.8; Daishin agar, 8 g L−1). In vitro culture was performed in a growth cabinet (Elbanton) at 28°C under a 16:8 h light/dark photocycle and 180 µmol m2‐1 s−1 light intensity. Parasponia andersonii plantlets were grown for 10 weeks in soil, in a greenhouse at 28°C and 85% humidity and under a 16:8 h light/dark photocycle. Plants were watered two times a week with water or nutritive solution (NH4, 1 mmol L−1; K, 5.9 mmol L−1; Ca; 2.7 mmol L−1; Mg, 0.8 mmol L−1; NO3, 9 mmol L−1; SO4, 1.9 mmol L−1; P, 1.1 mmol L−1; Fe, 15 µmol L−1; Mn, 5 µmol L−1; Zn, 5 µmol L−1; B, 3 µmol L−1; Cu, 0.5 µmol L−1; Mo, 0.5 µmol L−1, pH 5.8). An additional 100 mL of NH4NO3 (10 mm) (pH 5.8) was provided every week to each pot to abolish eventual symbiotic associations of P. andersonii with bacteria naturally present in the soil. Axillary shoots suppression was performed manually, one time a week, on wild‐type PanWU1‐14 and control plants PanCtr‐44 (van Zeijl et al., 2018). Internode diameters were measured using a digital caliper (#100436; Deubba GmbH, Merzig, Germany). Secondary growth assays were performed twice.

RNA isolation from P. andersonii epicotyls and stems

Parasponia andersonii epicotyl or stem RNA isolation was performed as described by van Velzen et al., (2017). The detailed experimental procedure for RNA isolation is provided in Methods S1.

qRT‐PCR gene expression analysis

Full‐length cDNAs were synthesized from 1 µg of RNA using SuperScript II Reverse Transcriptase kit (Thermo Fisher, Waltham, MA, USA) in the presence of Ribolock RNase Inhibitor (Thermo Fisher). qRT‐PCR analysis was performed on 1 µl of five‐fold diluted cDNA templates using iQ SYBR Green Super‐mix (Bio‐Rad, Hercules, CA, USA) and a CFX Connect Optical Cycler (Bio‐Rad) in accordance with the manufacturer’s instructions. Cycling conditions were set as: one pre‐incubation cycle (95°C, 3 min) and 40 amplification cycles (denaturation, 95°C, 15 s; hybridization–elongation, 60°C, 30 s). For the melting curve, conditions were set as: (denaturation, 95°C, 10 s; hybridization, 65°C, 5 s; denaturation until 95°C with 0.5°C incrementation). Cycle thresholds and primer specificities were determined using cfx maestro (Bio‐Rad). PanELONGATION FACTOR1α (PanEF1α) was used as a reference gene to normalized target gene expressions (van Zeijl et al., 2018). The qRT‐PCR data resulted from the analysis of three biological replicates and two technical replicates. Information concerning primers used for qRT‐PCR gene expression analysis is provided in Table S2.

RNA in situ hybridization

RNA in situ hybridizations were performed as described by Liu et al., (2019) using the Invitrogen ViewRNA ISH Tissue1‐Plex Assay kits (Thermo Fisher) and in accordance with the manufacturer’s instructions. RNA in situ hybridization experiments were repeated three times. A detailed experimental procedure is provided in Methods S1.

Promoter:GUS:terminator reporter fusion construction

Sequence information for PanNOOT1 (PanWU01x14_292800) promoter and terminator regions were retrieved from the P. andersonii genome (www.parasponia.org) (van Velzen et al., 2018). Promoter and terminator sequences were amplified using High Fidelity Phusion polymerase (Thermo Fisher) and cloned by golden gate cloning (Engler et al., 2014). For the PanNOOT1 promoter, three DNA fragments of 1.171, 1.401 and 0.820 kb, respectively, were cloned into level −1 pAGM1311 universal acceptors using BsaI and assembled into level 0 pICH41295 acceptor (PROM + 5’UTR) using BpiI. For the PanNOOT1 terminator, a DNA fragment of 0.973 kb was cloned into level 0 pICH41276 acceptor (3’UTR + TER) using BpiI. Information related to the primers used for the promPanNOOT1:GUS:PanNOOT1ter construction is provided in Table S3. The PanNOOT1 promoter (3.385 kb), uidA (2.001 kb) from level 0 pICH75111 module and the PanNOOT1 terminator (0.973 kb) were assembled into level 1 pICH47751 acceptor using BsaI, resulting in a 6.359 kb construction. The construction was combined with level 1 pICH54011:dummy 1, pICH47742:promNOS:HYGROMYCIN:NOSter and pICH41766:level 2 end‐linker 3 parts into level 2 pICSL4723 acceptor using BpiI. The level 2 construction was finally introduced into Agrobacterium tumefaciens strain AGL1 (Lazo et al., 1991) for subsequent plant transformations.

Promoter:GUS:terminator gene expression pattern

Histochemical GUS staining was performed as described in Pichon et al. (1992). Briefly, stem samples were vacuum infiltrated for 90 min (approximately 500 mmHg) in X‐gluc staining buffer (50 mm phosphate buffer (pH 7.2), 1 mm potassium ferricyanide, 1 mm potassium ferrocyanide, 0.1% (w/v) SDS, 1 mm EDTA and 1.25 mm 5‐bromo‐4‐chloro‐3‐indolyl‐beta‐d‐GlcA containing cyclohexylammonium salts) and incubated at 37°C for 20 h, under darkness. Samples were fixed in 50 mm phosphate buffer (pH 7.2), 1% (v/v) glutaraldehyde and 4% (v/v) formaldehyde for 15 min under vacuum (approximately 500 mmHg). The GUS experiments were repeated two times for 7–8‐week‐old transgenic plants and performed on six independent transformed‐lines.

Agrobacterium tumefaciens‐mediated transformation of P. andersonii, CRISPR‐Cas9 genome edition strategy and CRISPR‐Cas9 mutants genotyping

These procedures were performed in accordance with the experimental procedures described by van Zeijl et al., 2018 and Wardhani et al., 2019. Details of the experiments are also provided in Methods S1.

Technovit sections of P. andersonii woody stem

Parasponia andersonii woody‐stem sections were performed using Technovit 7100 (Kulzer GmbH, Wehrheim, Germany) in accordance with the manufacturer's instructions. Minor modifications were applied to the original procedure to preserve the structural integrity of the different wood tissues. As a major modification, wood samples were softened using an aqueous solution of ethylenediamine, 4%. A detailed experimental procedure is provided in Methods S1.

Phloroglucinol staining of P. andersonii stem lignins

The staining of P. andersonii stem lignins was performed using a phloroglucinol‐HCl solution as described by Pradhan Mitra and Loqué (2014). Phloroglucinol stainings were repeated twice. A detailed experimental procedure is provided in Methods S1.

RNA‐seq and data analysis

All of the internode samples used in the RNA‐seq analysis consisted of three biological replicates. Single end 50‐bp reads were sequenced using a BGISEQ‐500 sequencing system (BGI, Shenzhen, China), yielding an average of 4.5 Gbp per sample. All reads were deposited at the European Nucleotide Archive (www.ebi.ac.uk/ena) under accession number PRJEB37036. Gene expression was quantified by pseudo‐aligning the reads to the P. andersonii coding sequences (www.parasponia.org) (van Velzen et al., 2018) using kallisto, version 0.43 (Bray et al., 2016) with default parameters. Tissue‐dependent and genotype‐dependent genes were identified with the model testing framework as implemented in sleuth (Pimentel et al., 2017). Briefly, for every gene, a likelihood ratio test was performed comparing a model with only an intercept term versus a model with a term for either tissue type or genotype. From this, genes with a corrected P ≤ 0.01 were selected as varying significantly over tissue or genotype. Subsequently, distinct expression profiles were identified by performing K‐means clustering on the significantly varying genes. Several values of k were tested and, for both tissue‐dependent and genotype‐dependent genes, no more than two visually distinct expression profiles could be identified. The code used for RNA‐seq analysis is available at https://github.com/holmrenser/parasponia_code. Parasponia andersonii internode transcriptomic data have been integrated into an interactive website (https://parasponia.plantgenie.org) constructed using genie‐sys (Mannapperuma et al., 2019). These transcriptomic data will be integrated into the next version of the PlantGenIE resource (Sundell et al., 2015).

Parasponia andersonii transcription factor identification and analysis

Parasponia andersonii proteins were used as queries in blastp searches against A. thaliana protein sequences in The Arabidopsis Information Resource 10 (TAIR) (https://www.arabidopsis.org) database to obtain the closest A. thaliana homolog (E‐value ≤ 1 × 10–10). The set of A. thaliana transcription factors from Plant TFDB 5.0 was used as the reference database to annotate and assign P. andersonii transcription factors to a transcription factor family (http://planttfdb.cbi.pku.edu.cn/index.php?sp=Ath) (Jin et al., 2014).

AUTHOR CONTRIBUTIONS

KM and RG conceived the project and designed the experiments. KM performed the histological kinetic of P. andersonii cambium establishment. DS and KM performed the qRT‐PCR gene expression analysis. OK and DS performed the PanNOOT1 RNA in situ hybridization. KM cloned the promPanNOOT1:GUS construct and performed the histological analysis. TVDM and KM genotyped the promPanNOOT1:GUS transgenic plants. DS generated the Pannoot1 CRISPR‐Cas9 loss‐of‐function mutants with the help from FB. KM and DS performed the Pannoot1 secondary growth characterization. KM performed the wood sections of P. andersonii and analyzed the images. KM performed the lignin deposition analysis. DS and YZ collected the material and isolated the RNA for the transcriptomic analysis. RH performed the transcriptomic analysis and the statistical analysis. CM and NRS conceived the P. andersonii interactive website and integrated P. andersonii transcriptomic data into PopGenIE. DS, RH, ZY, OK, RG and KM analyzed the data. KM, DS, RH and ZY conceived the figures. KM, DS and RG wrote the manuscript. DS, RG and KM revised the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Figure S1. Correlation between P. andersonii cambium establishment and shoot development.

Figure S2. PanNOOT1 gene expression profile.

Figure S3. PanNOOT1 promoter GUS expression pattern in P. andersonii stem.

Figure S4. Description of the Pannoot1 CRISPR‐Cas9 mutants.

Figure S5. Illustrations of the different genotypes used for the secondary growth assay.

Figure S6. The phellogen of P. andersonii.

Figure S7. Pannoot1 A5 stems secondary growth assay at 20 weeks.

Figure S8. Stem, xylem and phloem areas are reduced in Pannoot1.

Figure S9. Parasponia andersonii internode sections corresponding to the samples used in the transcriptomic analysis.

Figure S10. Parasponia andersonii PanNOOT1 RNA‐seq expression profile in PanCtr‐44 and Pannoot1 A5 and A10 internodes.

Figure S11. Top ten enriched Gene Ontology terms.

Figure S12. Tissue‐ and genotype‐dependent differentially expressed transcription factors.

Figure S13. Lignin deposition and heatmap of putative lignin biosynthesis‐related genes expression in Pannoot1 internodes.

Figure S14. Tissue‐ and genotype‐dependent expression of key secondary growth‐related genes.

Figure S15. Gene expression profiles of Populus tremula and Betula pendula BOP genes in stem tissues.

Figure S16. A model for PanNOOT1 stem secondary growth regulation.

Click here for additional data file.

Table S1. Secondary growth‐related orthogroups. Orthogroup numbering according to van Velzen et al. (2018).

Click here for additional data file.

Table S2. Primers used for qRT‐PCR.

Table S3. Primers used for promPanNOOT1:GUS:PanNOOT1ter golden gate cloning.

Click here for additional data file.

Dataset S1. Phylogenetic reconstruction analysis of putative P. andersonii secondary growth‐related genes orthologs.

Click here for additional data file.

Dataset S2. Lists of the genotype‐ and tissue‐dependent differentially expressed transcripts and transcripts per million values.

Click here for additional data file.

Dataset S3. Lists of differentially expressed transcription factors.

Click here for additional data file.

Methods S1. Enhanced experimental procedures.

Click here for additional data file.