Auxin signalling at the vascular cambium coordinates the development of wood fibres in a noncell-autonomous manner
To comprehensively analyse the developmental process of the secondary xylem in Populus tomentosa, we conducted cross-sectional observations of stem segments from 3-month-old plants in different internodes. We observed that the fully formed ring-shaped DSX occurred in the 6th internode (Supplementary Fig. 1a). Within the DSX region, newly differentiated vessels and fibres underwent cellular expansion, with cell sizes significantly larger than those of cambial cells that are not yet lignified (Supplementary Fig. 1a). Remarkably, a fixed number of developing fibre (DF) cell layers (3-5), were present from the examined 6th to 15th internodes, thus signifying a stable and easily identifiable DSX phenotype (Supplementary Fig. 1b).
To elucidate the potential impact of local auxin signalling on wood fibre development at the DSX, we introduced the IAA9m gene encoding an auxin-resistant variant of the IAA9 protein, which is a transcriptional repressor of auxin signalling21, into P. tomentosa driven by the SND1-A1 promoter. The SND1-A1 gene is reported as a SND1 homologue in poplar and is preferentially expressed at DSX9, 33. Notably, the transgenic lines displayed a marginal reduction in plant height, while internode number and stem diameter remained unaltered compared to the wild-type (WT) counterparts (Supplementary Fig. 2a-f). Furthermore, consistent numbers of DF cell layers were observed in the 7th, 8th, and 9th internodes of both WT and transgenic plants (Fig. 1a, b, Extended Data Fig 1a, b). These results show that local auxin signalling at DSX does not exhibit significant effects on wood fibre development.
Given the high accumulation of auxin in the cambial zone of poplar stems17, 19, we hypothesized noncell-autonomous modulation of cambium-localized auxin signalling impacts wood fibre development at the DSX. To test this hypothesis, auxin signalling was attenuated by expressing IAA9m driven by the promoter of poplar WOX4a, a well-known vascular cambium-specific gene23-25. The transgenic lines displayed substantial reductions (10.2-34.2%) in plant height, maintained consistent internode counts, and had slightly reduced stem diameters, compared with the WT (Supplementary Fig. 2g-l). We further selected two transgenic lines (#2 and #3) with high IAA9 expression and stunted growth for detailed analysis. Stem cross-sections revealed a consistent reduction in the number of DF cell layers within all analysed internodes (7th, 8th and 9th) of WOX4apro:IAA9m lines in comparison to their respective counterparts in the WT (Extended Data Fig. 1a, b). Notably, in contrast to the 3-5 cell layers present at the 8th internode of WT and SND1-A1pro:IAA9m plants, WOX4apro:IAA9m lines displayed only 1-3 DF cell layers (Fig. 1a, b). These results suggest diminished DF cell layers when auxin responsiveness was impaired in the vascular cambium.
We assessed wood fibre cell dimensions to probe the influence of cambium-localized auxin signalling on wood fibre development. Isolated lignified fibres (LFs) from SND1-A1pro:IAA9m stems displayed cells of comparable length and width to those of the WT (Fig. 1c-e, Extended Data Fig. 1c-e). Conversely, compared to the WT, WOX4apro:IAA9m LFs exhibited significantly diminished dimensions, with reductions of 20.5-22.7% in length and 22.3-28.6% in width (Fig. 1c-f, Extended Data Fig. 1c-e). These results highlight that restricted auxin signalling within the vascular cambium constrains the expansion of wood fibre cells. Additionally, we investigated vessel morphology in the IAA9m transgenic lines. Both SND1-A1pro:IAA9m and WOX4apro:IAA9m lines displayed vessel sizes akin to the WT (Supplementary Fig. 3a-d). Taken together, these findings reveal the noncell-autonomous role of auxin signalling from the vascular cambium in DSX wood fibre development.
PLT5a/b specifically respond to auxin signals in the vascular cambium and regulate wood fibre development
To unravel the noncell-autonomous regulatory mechanism of auxin signalling in wood fibre development, we reasoned that auxin signals at the vascular cambium may trigger specific downstream factors, acting as intermediaries to regulate wood fibre development. Initially, we sought to identify auxin-inducible transcription factor (TF) genes specifically expressed in the vascular cambium. In previous work, we employed 35Spro:IAA9m transgenic poplar plants with constitutively inhibited auxin signalling and detected 2170 downregulated genes21. Among these, 767 genes were allocated to the cambium-specific gene clusters identified in the AspWood datasets (Extended Data Fig. 2a)26. Gene Ontology (GO) analysis indicated the enrichment in categories linked to auxin response and DNA binding (Extended Data Fig. 2b). Notably, 27 TF genes, including 15 TFs identified on the list of vascular cambium-specific TFs derived from P. trichocarpa cell-type transcriptome analysis, were downregulated by IAA9m overexpression (Extended Data Fig. 2c)25. These TFs belong to the AP2, DOF, HD-ZIP, bHLH, and E2F families, encompassing two homologous proteins (termed PLT5a and PLT5b) from the AP2 TF subfamily (Extended Data Fig. 2d). A phylogenetic analysis established PLT5a and PLT5b as counterparts to the Arabidopsis PLT5 protein in poplar (Extended Data Fig. 2e). Amino acid sequence alignment revealed high similarity of the AP2 domains between P. tomentosa PLT5 proteins and their Arabidopsis counterpart (Supplementary Fig. 4a), suggesting conserved functionality as TFs. Furthermore, previous studies in Arabidopsis have demonstrated that a suite of PLT proteins respond to auxin signalling and serve as pivotal regulators governing root cell division and differentiation27-30. Thus, we focused on investigating the role of PLT5a/b proteins in wood fibre development.
In comparison to the WT, the expression of PLT5a and PLT5b was significantly reduced by 37.6% to 53.5% in the WOX4apro:IAA9m lines, while it remained unchanged in the SND1-A1pro:IAA9m lines (Fig. 2a), suggesting a specific response of PLT5a/b to auxin signalling in the poplar stem cambium. Notably, exogenous auxin treatment robustly induced the expression of both PLT5 genes (Fig. 2b), confirming their auxin responsiveness. Detailed sequence analysis revealed the presence of at least seven core auxin response elements (AuxREs) within the promoter of PLT5a genes (Extended Data Fig. 3a). Furthermore, the AuxRE-containing fragments from the PLT5a promoter were enriched in chromatin immunoprecipitation (ChIP) assays (Extended Data Fig. 3a). Electrophoretic mobility shift assays (EMSAs) confirmed the binding capacity of these AuxRE-containing promoter fragments by AUXIN RESPONSE FACTOR (ARF) 5.1 (Extended Data Fig. 3b), a crucial auxin signalling transcription factor expressed in the poplar cambium 21. Impressively, ARF5.1 effectively activated luciferase reporter expression driven by AuxRE-containing fragments I and III of the PLT5a promoter in effector-reporter assays (Extended Data Fig. 3c, d). This activation was abrogated when AuxREs were disrupted by site-directed mutagenesis, underscoring the AuxRE-dependent regulation of PLT5 expression within the context of auxin signalling. Collectively, these findings strongly imply that PLT5 may represent a direct auxin-responsive gene targeted by the ARF TF.
The AspWood and laser capture microdissection-based cell-type transcriptome data demonstrated PLT5s' predominant expression within the vascular cambium of P. tremula and P. trichocarpa (Supplementary Fig. 4b, c)25-26. In line with this, in situ hybridization assays revealed selective enrichment of PLT5 mRNAs in the vascular cambium of P. tomentosa (Fig. 2c, d). Furthermore, we established YFP reporter lines driven by the native PLT5a promoter, resulting in fluorescent signals primarily localized to the cambial zone (Fig. 2e, f). In summary, we identified a pair of PLT5 genes encoding vascular cambium-specific auxin-responsive transcription factors in poplar.
Using CRISPR/Cas9 genome editing, we generated loss-of-function mutants for PLT5a and PLT5b, employing four sgRNA-targeted sites (T1-2 for PLT5a; T3-4 for PLT5b) to create double knockout mutants. Two selected transgenic lines (plt5a plt5b #4 and #5) with nucleotide modifications were further studied (Supplementary Fig. 5a). To enhance PLT5 expression specifically in the vascular cambium, we utilized the vascular cambium-specific WOX4a promoter to drive PLT5a expression (Supplementary Fig. 5b, c). The plt5a plt5b mutant showed a significant reduction in plant height, internode count, and stem diameter (Supplementary Fig. 5d-g). In contrast, the WOX4apro:PLT5a lines exhibited increased plant height and stem diameter but maintained similar internode counts to those in the WT. Furthermore, the plt5a plt5b mutants showed reduced DF cell layers in the 7th to 9th internodes (Extended Data Fig. 4a, b), resembling the DSX-deficient phenotype of WOX4apro:IAA9m plants (Extended Data Fig. 1a, b). In plt5a plt5b line #4, the 8th internode exhibited only 1-3 DF cell layers compared to the WT (Fig. 2g, h). Conversely, the WOX4apro:PLT5a lines had increased DF cell layers (4-6; Fig. 2g, h, Extended Data Fig. 1a, b). These findings suggest a positive role for PLT5 genes in maintaining DF cell layers within DSX.
Divergent effects on fibre cell expansion were observed in the plt5a plt5b and WOX4apro:PLT5a lines. The length of LFs decreased by 29.2% to 30.4% in plt5a plt5b mutants (#4 and #5) but increased by 27.5% to 37.1% in WOX4apro:PLT5a lines (#1 and #2) compared to those in the WT (Fig. 2i, j, Extended Data Fig. 4c, d). Similarly, the plt5a plt5b and WOX4apro:PLT5a lines exhibited reduced and increased LF widths, respectively (Fig. 2k, Extended Data Fig. 4e). Notably, no alterations in vessel cell size were detected between the WT and transgenic lines (Supplementary Fig. 6a-d), highlighting the specific influence of PLT5s on wood fibre development in poplar.
Mobile PLT5a proteins mediate noncell-autonomous auxin signalling in wood fibre development
Given that impaired auxin responsiveness in the vascular cambium and PLT5 loss-of-function both led to similar fibre phenotypes, we propose PLT5 as an auxin-regulated mediator of fibre development. YFP-tagged PLT5a reporter lines revealed nucleus-localized PLT5 proteins in both cambial and DSX cells (Fig. 3a, b, Supplementary Fig. 7a), suggesting a shift of PLT5 proteins from the vascular cambium to DSX cells. We confirmed their mobility using a triple-YFP tag that hindered cell-to-cell movement of mobile proteins in Arabidopsis31-32. PLT5a fused with triple-YFP (PLT5apro:PLT5a-YFP ×3) showed fluorescence primarily in cambial cells and rarely in DSX cells (Fig. 3a, b), supporting PLT5 protein mobility over short distances.
PLT5apro:PLT5a-YFP lines displayed increased plant height and stem thickness at lower internodes, while the internode number remained unchanged compared to those in the WT (Supplementary Fig. 7b-d). Similar results were observed for PLT5apro:PLT5a-YFP×3 lines in terms of plant height, internode number, and diameter (Supplementary Fig. 7b, c, e). Stem cross-sections revealed that PLT5apro:PLT5a-YFP lines had increased DF cell layers and fibre size (Supplementary Fig. 7e-i), paralleling the fibre phenotypes observed in WOX4apro:PLT5a lines (Fig. 2g-k). In contrast, the PLT5apro:PLT5a-YFP×3 lines showed no changes in DF cell layers or fibre size (Supplementary Fig. 7e-i), highlighting the importance of PLT5 protein mobility.
To establish the link between PLT5 and auxin-dependent wood fibre development, we introduced functional single- or triple-YFP fusions of PLT5 into WOX4apro:IAA9m lines (Supplementary Fig. 8a). Both fusions slightly increased plant height without affecting internode number and stem diameter (Supplementary Fig. 8b-g). PLT5a-YFP restored DF cell layers in WOX4apro:IAA9m from 2.0 to 3.7, even surpassing the WT levels (average 3.1; Fig. 3c, d). However, this rescue was absent when the immobile triple-YFP fusion of PLT5awas introduced into the WOX4apro:IAA9m background (Fig. 3c, d). Similarly, the single-YFP fusion of PLT5a fully restored the deficient cell expansion of WOX4apro:IAA9m fibres both in length and width (Fig. 3e-g). While partial restoration (28.6% for length and 33.1% for width) occurred with PLT5a-YFP×3 in WOX4apro:IAA9m lines, these fibres remained significantly shorter and thinner than the WT fibres (Fig. 3e-g). These differential complementation patterns emphasize the role of mobile PLT5 proteins in the noncell-autonomous auxin signalling effect on wood fibre development.
PLT5 coordinates wood fibre development by regulating cell wall thickening
To investigate the mechanism governing wood fibre development mediated by PLT5, we employed phloroglucinol-HCl staining to assess lignification in xylem cell walls. Notably, a substantial reduction in lignin deposition in the DF cell wall in DSX was observed in both the WOX4apro:PLT5a and PLT5apro:PLT5a-YFP lines (Supplementary Fig. 9a, b), indicating that cambium-specific expression of PLT5 inhibits SCW lignification. Scanning electron microscopy (SEM) analyses further revealed that in plt5a plt5b mutant stems, the cell wall thickness of the 1st to 5th DF cell layers in the 8th internode significantly increased by 17.7% to 83.3% compared to those in the WT (Fig. 4a, b, Extended Data Fig. 5a). Conversely, in the WOX4apro:PLT5a lines, the corresponding DF cell layers exhibited a reduction in cell wall thickness by 31.9% to 54.4% compared to those in the WT (Fig. 4a, b, Extended Data Fig. 5a). However, knockout or overexpression of PLT5 at the cambium did not alter the cell wall thickness of LFs (Supplementary Fig. 11a, b). Moreover, expression of PLT5a-YFP under its native promoter yielded a phenotype similar to that observed in WOX4apro:PLT5a lines (Fig. 4a, b, Supplementary Fig. 10a, b, 11c, d). However, expression of the triple-YFP fusion of PLT5a(PLT5apro:PLT5a-YFP×3) did not delay the thickening process of DF cell walls (Supplementary Fig. 10a, b). Thus, these results indicate that the suppressed thickening of DF cell walls by PLT5 relies on the mobility of their proteins.
We further explored the influence of auxin signalling on DF cell wall thickening. SND1-A1pro:IAA9m lines showed no discernible variations in cell wall thickness for each DF cell layer compared to those in the WT (Extended Data Fig. 5b, c). In contrast, within the initial five DF cell layers of WOX4apro:IAA9m lines, the cell wall was significantly thicker, ranging from 0.49 to 1.01 times the thickness of the WT cell wall (Extended Data Fig. 5b, c). No noticeable difference in cell wall thickness was detected for LFs between the WT and WOX4apro:IAA9m lines (Supplementary Fig. 11e, f). Importantly, the attenuation of DF cell wall thickening induced by WOX4apro:IAA9m could be reversed by the single-YFP fusion of PLT5a, but not the triple-YFP fusion (Fig. 4c, d, Extended Data Fig. 6d). These results reveal that mobile PLT5 proteins mediate noncell-autonomous inhibition of auxin signalling from the vascular cambium on DF cell wall thickening.
Based on variations in cell length and toluidine blue staining, isolated fibres were categorized into four classes (Extended Data Fig. 6a). Classes I and II comprised minimally stained DFs undergoing elongation. Class III fibres displayed slight staining, indicating early lignification initiation, while Class IV encompassed highly lignified fibres. An increased proportion of Class III and IV fibres was observed in the WOX4apro:IAA9m and plt5a plt5b lines (Supplementary Fig. 12a, b), indicating accelerated wood fibre lignification. In contrast, the developing and mildly lignified fibres (Classes I, II and III) in WOX4apro:PLT5a and PLT5apro:PLT5a-YFP lines exhibited higher percentages than those in WT (Supplementary Fig. 12b, c). The accelerated wood fibre lignification induced by WOX4apro:IAA9m could be restored by the cambial-specific expression of PLT5a-YFP (Supplementary Fig. 12d). Moreover, fibres belonging to Classes I and II of the WOX4apro:IAA9m and plt5a plt5b lines showed similar cell lengths to those in the WT, while those of Classes III and IV were notably shorter (Extended Data Fig. 6a, b). These findings suggest that the impairment in wood fibre cell expansion resulting from disrupted auxin signalling in the vascular cambium and plt5 knockout may coincide with SCW deposition.
To establish the link between fibre cell expansion and cell wall thickening coupled by auxin-PLT5 module, we quantified SCW thickness and cell area of DFs along radial files for modeling fibre cell growth. The thickening of fibre cell wall accompanied with the depth of fibre cell layers was fitted by a logistic growth equation (Extended Data Fig. 7a, b, Supplementary Fig. 13a, b). An inflection point (tI) emerges when the rate of SCW thickening reaches its maximum value. Compared to the WT (tI = 5.6), the plt5a plt5b mutants attain its maximum rate of SCW thickening in a much shorter time (tI = 1.8) while WOX4pro:PLT5a transgene with up-regulation of PLT5 expression in the vascular cambium delays the occurrence of its inflection point (tI = 7.6; Extended Data Fig. 7a). Therefore, vascular cambium-specific PLT5 gene negatively regulates the speed of SCW thickening at the early stages of wood fibre development. Similar to plt5a plt5b mutants, the WOX4pro:IAA9m transgene that compromises auxin signalling in the vascular cambium accelerates the occurrence of inflection point (tI = 3.0; Extended Data Fig. 7b). The single-YFP fusion of PLT5a protein (PLT5apro:PLT5a-YFP) can empowers the WOX4pro:IAA9m line to delay the inflection point (tI = 7.1; Extended Data Fig. 7b). These results demonstrate the PLT5-mediated suppression of vascular cambium-derived auxin signaling on the rate of SCW thickening in wood fibers, providing a critical link between these processes in fiber cell development.
We further evaluated the impact of cell wall thickening on wood fibre cell growth. The overall trajectories of fibre cell expansion indicated by the lumen area of a single fibre cell is partitioned into independent and dependent components (Fig. 4e). As expected, the independent component follows a growth equation, but due to the negative influence of cell wall thickening, the fibre cell growth displays a convex curve, and the emergence of the apex depends on genotypes (Fig. 4e). Compared to the WT, fibre cell expansion of plt5a plt5b is negatively affected by its cell wall thickening at the earlier stages and to a larger extent (Fig. 4e). In contrast, the WOX4pro:PLT5a line displays the increasing cell growth of fibres independent of the restriction resulting from cell wall thickening (Fig. 4e). Consistent with the plt5a plt5b mutant, an increasing negative effect of cell wall thickening on cell expansion is observed in the WOX4pro:IAA9m fibres (Fig. 4e). The introduction of PLT5apro:PLT5a-YFP encoding YFP-fused mobile PLT5a proteins compromises the dependent component of cell wall thickening while recovers the independent cell growth of fibres (Fig. 4e). These modeling data reveal that the auxin-PLT5-dependent coordination of cell expansion in fibres may result from their negative regulation on cell wall thickening at early stages of wood fibre development.
PLT5a directly represses SND1 expression to regulate SCW deposition in wood fibres
Given the influence of PLT5 on wood fibre SCW deposition, we explored the expression of key genes responsible for SCW biosynthesis, including cellulose synthase (CesA) and caffeic acid O-methyltransferase (COMT), along with critical transcription factor genes. Quantitative PCR (qPCR) analyses revealed elevated expression of these genes in WOX4apro:IAA9m and plt5a plt5b mutants, whereas significant suppression was observed in the WOX4apro:PLT5a line, compared to those in the WT (Supplementary Fig. 14a-d). Similarly, we observed modulated expression of all SND1 homologue genes, which serve as pivotal switches in the fibre-specific SCW transcription regulatory network (TRN), among these transgenic lines (Extended Data Fig. 8a, b), consistent with their DF cell wall phenotypes. These findings indicate that PLT5-mediated auxin signalling from the cambium inhibits the fibre-specific SCW TRN. Notably, unlike the triple-YFP fusion, the single-YFP fusion of PLT5a did not attenuate expression of these SND1 homologue genes (Extended Data Fig. 8b), emphasizing the vital role of PLT5 protein mobility in governing SCW deposition in wood fibres.
Previous studies have shown that Arabidopsis PLT5 shares a consensus DNA binding motif sequence with its homologue AINTEGUMENTA (ANT)33-34. Sequence analysis revealed the presence of several core motifs of potential PLT5-binding sites within the promoters of four SND1 genes from the P. tomentosa genome (Fig. 5a, Extended Data Fig. 9a-c). ChIP assays confirmed the association of PLT5 proteins with at least one predicted binding site in each SND1 promoter (Fig. 5a, Extended Data Fig. 9a-c). The binding capacity of PLT5 was confirmed through a shift in labelled probes containing Region IV of the SND1-A1 promoter in EMSA experiments (Fig. 5b). Effector-reporter assays further demonstrated the PLT5-dependent regulation of SND1 expression. Cotransformation of PLT5a significantly suppressed the expression of the luciferase reporter driven by the investigated binding region of each SND1 promoter (Fig. 5c, d, Extended Data Fig. 9d-i). Conversely, mutagenesis of these binding sites abolishes the transcriptional repression of PLT5aon SND1 genes (Fig. 5c, d, Extended Data Fig. 9d-i). These results indicate that PLT5 proteins directly repress the expression of SND1 homologues by binding to specific DNA motifs in their promoters.
To establish the genetic link between PLT5-mediated auxin signalling and SND1-driven SCW deposition in wood fibres, we performed genetic complementation of PLT5-induced fibre phenotypes using the SND1-A1 gene (Supplementary Fig. 15a). Overexpression of SND1-A1 resulted in reduced plant height and stem diameter in both the WT and WOX4apro:PLT5a backgrounds, while the internode number remained unaffected (Supplementary Fig. 15b-f). Ectopic SCW deposition occurred in the parenchyma cells of SND1-A1-overexpressing lines (Supplementary Fig. 15g), consistent with previous reports10. Next, we investigated the impact of SND1-A1 on DF phenotypes, observing that the overexpression of SND1-A1 led to a reduction in DF thickness in WOX4apro:PLT5a lines by 3-5 cell layers, comparable to those in the WT (Fig. 5e, f, Extended Data Fig. 10a). The increased fibre length and width observed in WOX4apro:PLT5a lines were restored by the introduction of SND1-A1 (Extended Data Fig. 10b-d). Similarly, recovery was also observed for the decreased proportion of highly lignified fibres caused by PLT5 (Extended Data Fig. 10e). Additionally, quantitative analyses further demonstrated that SND1-A1 overexpression resulted in increased DF cell wall thickness compared to those in the WT, and effectively restored the altered DF cell wall thickness caused by WOX4apro:PLT5a (Fig. 5g, h, Extended Data Fig. 10f). Taken together, these findings conclusively establish the role of SND1-A1 as a pivotal downstream mediator of the regulatory interplay between the auxin-PLT5 cascade and wood fibre development in poplar.