Microtubules play an important, but not vital role, in MFA determination and reorientation during wood formation.
MFA, the angle at which cellulose microfibrils are deposited in the secondary cell wall, is an important characteristic of reaction wood formation. In angiosperms, tension wood features very low MFA while opposite wood reportedly presents larger angles (Almeras and Clair 2016). At the upper side of angiosperm branches, the alignment of cellulose microfibrils at low angles gives support to non-vertical stems by opposing tensional forces, whereas in gymnosperm compression wood, large MFA acts by opposing compression forces (Bamber 2001). While many studies focus on understanding tension wood, active re-modelling of opposite wood was also implied by transcriptomic studies, which found substantial transcriptional changes occurring in opposite wood compared to normal wood (Chen et al. 2015; Groover 2016). In our experiments, both MFA and MTA were lower in tension wood when compared to opposite wood in both eudicots, which is consistent with expectations. By contrast, normal wood usually displays intermediate MFA values between tension wood and opposite wood (Almeras and Clair 2016). We observed this relationship for normal wood in eucalypt and pine; however, in our study poplar normal wood presented the lowest values for both MTA and MFA. This discrepancy might relate to the species and the effects glasshouse conditions. Environments conditions were favorable for fast growth and free of forces such as wind that led to poplar stems growing longer and thinner relative to other species as they competed for light, altering load dynamics within the stem. Consequently, this fast competitive stem growth may have induced tension near the base of vertical stems leading to the observed lower-than-expected MTA and MFA values.
Tubulin genes are highly expressed in reaction wood. In tension wood, for example, expression levels of some tubulin genes can increase as much as 3.4-fold in comparison to opposite wood (Qiu et al. 2008), while in Arabidopsis, genes related to microtubule formation were downregulated under microgravity conditions which corresponded with suppression of reorientation processes (Kato et al. 2022). It seems that transcriptional, rather than translational regulation was more influential in determining the composition of tubulin isotypes in woody tissue, suggesting isotypes may play a role in this aspect of microtubule function (Hu et al. 2015). While these results indicate that there are more tubulin proteins present in reaction wood cells, exactly how the quantity of tubulin might determine microtubule assembly and reorientation is unknown. Here, high correlations between MTA and MFA were observed across all studies, which also correlated with cell wall thickness. Loss of an organised microtubule array via oryzalin treatment perturbed the reorientation of cellulose microfibril deposition in in vitro cultured wood cells, while stabilisation of microtubule arrays in planta presumably halted re-alignment of cellulose synthesis in reaction wood, resulting in cell wall changes. Taken together these finding suggest an important, but not essential role, for cortical microtubules arrays in the re-alignment of cellulose microfibril deposition.
Using microtubule-disrupting and -stabilising drugs, we documented that microtubule dynamics are important, but not essential, for reorientation of microfibril deposition during wood formation. Here, we showed lower MFA in fibres and tracheids cultured in vitro in the presence of oryzalin, although these results were not statistically significantly different from control in vitro stem segments. A move to from in planta to in vitro conditions subjects stems to different concentrations of phytohormones, gravitropic stimuli, environmental cues and inter- and intracellular forces, leading to in vitro derived wood and xylem fibres with distinct changes in cell wall properties (Leitch and Savidge. 1995, Decou et al. 2020), which may have masked some of the changes in microtubule organisation and therefore, MFA. We therefore focused on injection of the microtubule-stabilising drug paclitaxel into live, whole trees undergoing reaction wood formation. Previous studies have shown that paclitaxel stabilisation of microtubules prevents their reorientation in the primary cell wall of developing cotton fibres (Seagull 1990). Here, higher MTA and MFA were observed in poplar wood fibre cells treated with paclitaxel when compared to untreated cells, consistent with the idea that microtubules stabilisation may have prevented reorientation of cellulose deposition in tension wood.
Microtubules guide the CSC in secondary cell walls (Watanabe et al. 2015) and microtubule reorganisation and CSC realignment are particularly important in response to environmental and developmental cues (Kesten et al. 2019; Schneider et al. 2017). Microtubule arrays undergo constant reorganisation in response to environmental cues. Changes in microtubule orientation and microfibril deposition also occur during tracheary element development in woody stems (Abe and Funada 2005) and bending (Bisgrove, 2008). In tension wood, low MFA is likely orchestrated by a shift of the cortical microtubule array from a transverse to a longitudinal alignment with the reverse taking place in opposite wood. We also document strong coalignment between MFA and MTA in wood fibres and tracheids across a number of states and species, in line with these expectations. Studies of cells with only primary cell walls suggest that microtubule reorientation can be controlled by mechanical stresses (Landrien and Hamant, 2013) and it is therefore likely that mechanical forces in the stem and gravitropic stimuli contribute to changes in MTA and consequently, MFA. Overall, this data supports the premise that microtubules arrays play an important primary role in MFA determination during secondary cell wall formation in woody stems in response to environmental and mechanical cues.
Microfibril orientation does not exclusively result from microtubule guidance and it can change due to interaction with adjacent microfibrils and non-cellulosic components (McFarlane et al. 2014). In primary cell walls of Arabidopsis root epidermal and cortical cells, drug-dependent microtubule disruption does not alter parallel microfibril deposition alignment locally, but microfibrils showed considerable variation in their orientation within the same cell (Baskin et al. 2004). These results suggest that cellulose deposition into well-ordered parallel patterns can be generated by a self-assembly mechanism with little or no microtubule influence and they seem to spontaneously align. In this respect, Emons (1994) first proposed a geometrical model to explain the self-ordering mechanism of cellulose microfibrils based on the density of CSCs in the plasma membrane, the distance between newly deposited cellulose microfibrils and the overall cell geometry, which has been further supported by other studies (Himmelspach et al. 2003; Sugimoto et al. 2001). Other models incorporated the notion of a scaffold of proteins and/or polysaccharides built on the plasma membrane, which is based on microtubule orientation (Baskin 2001). For example, Schneider et al (2017) found that microtubule-driven patterns become 'imprinted' in cell walls during the xylem vessel development to the point where they are sufficient to sustain the continued development of wall thickening, suggesting once the cell wall patterning is established, there does not appear to be a direct need for ordered microtubules. Recently, Chan and Coen (2020) reported a dual mechanism of CSC guidance in Arabidopsis leaves: autonomous CSCs follow trails generated by previous complexes that are overridden when a CSC encounters a microtubule. Therefore, microtubule guidance seems to be dominant, playing a role in reorientation of cellulose microfibril deposition, but an autonomous system can take over in the absence of microtubules to ensure cell wall formation continues, which is consistent with our observations.
Microtubule alignment changes are coupled with changes in cell wall thickness but not G-layer formation
Thicker secondary cell walls with higher cellulose content are normally associated with reaction wood formation (Almeras and Clair 2016) and transcriptional studies have shown upregulation of cellulose synthase genes in these tissues (Li et al. 2013, Chen et al. 2015), demonstrating increased CSC activity or abundance in cells during reaction wood development. Throughout the diversity of samples and treatments we observed, we consistently documented an inverse relationship between microtubule bunding and reaction wood cell wall thickness. Eucalypt and pine samples both showed significantly thicker cell walls and decreased microtubule bunding in induced reaction wood compared to opposite wood, while neither of these properties were significantly different between poplar reaction and opposite wood. These findings are consistent with a recently proposed molecular model of MFA determination in trees forming reaction wood, which underlined the action of microtubule associated proteins (MAPs) in reorganizing the cortical microtubule array by promoting bundling of microtubules (Tobias et al. 2020). Specifically, MAP65 associates with microtubules to facilitate bundling (Meng et al. 2010) by acting as a spacer and structural element that promotes the bundling of microtubules that encounter each other at shallow angles (Tulin at al, 2012), while katanin, a microtubule severing enzyme, is involved in promoting microtubule array alignment by removing discordant microtubules (Deinum et al 2017, Yagi, et al 2021). Interestingly, MAP65 inhibits katanin binding in vitro leading to increasing bundling and crosslinking in the cortical microtubule arrays in primary cell walls (Burkart and Dixit, 2019). It is therefore possible that bundling may assist in microtubule reorientation through MAP65/katanin interactions where decreased bundling favors katanin action, supporting reorientation and facilitating parallel alignment. Consequently, less microtubule bundling would therefore be expected in cells undergoing cellulose deposition re-orientation and reaction wood formation, as observed here.
The G-layer is a chemically and morphologically distinct layer of the secondary cell wall characterised by extremely low MFA, a highly mesoporous matrix, and the ability to generate axial strain during tension wood formation in poplar (Clair et al. 2018). The development of a G-layer in poplar is a common response to gravitational stimuli and it is believed to rapidly generate tension stress (Abedini et al. 2015). Here, thicker cells wall in tension wood of poplar where not consistently observed, however G-layers were. While Eucalyptus globulus has been reported to produce typical G-fibres (Washusen et al. 2005), in our study this feature was not observed. While development of a G-layer is part of the process of secondary cell wall thickening during tension wood formation (Clair et al. 2018), our results suggest that microtubules do not substantially participate in G-layer formation, or at least microtubule (re)organisation is not crucial for G-layer development, since G-layer development was unaffected by paclitaxel-mediated microtubule stabalisation. Indeed, G-layer formation is regulated by a different set of transcription factors that affect secondary cell wall composition (Gorshkov et al. 2017). These G-layer transcription factors are more likely to be responsible for the development of distinct chemical properties in this layer and appear to be transcriptionally regulated via ethylene (Seyfferth et al, 2019).
Comparing wood formation across diverse species reveals conserved and distinct roles for microtubules
Dynamic microtubule reorganisation is an important component of reorganizing microfibril deposition during reaction wood formation. Computer simulation showed that branched nucleation increased array polarity (i.e. microtubule growth towards the same direction) and that changes in the mean branch angle increased the probability of the array to shift at least 20° from the initial transverse orientation (Eren et al. 2010). Indeed, some studies have demonstrated that branched nucleation followed by depolymerisation of mother microtubules is the main mode of array reorientation in plants (Nakamura 2015). This process appears to be facilitated by the action of both Augmin, a complex involved in recruiting gamma tubulin to initiate nucleation (Liu et al 2014) and prevents severing (Wang et al 2018), and katanin, a microtubule severing enzyme (Deinum et al 2017, Yagi, et al 2021). These findings suggest a model in which branching, followed by severing, could shift microtubule orientation and then maintain parallel alignment of the whole array. Here, we inferred new microtubule formation (nucleation) by the number of branching events observed in the microtubule arrays. In angiosperms, the amount of microtubule branching was lower in tension and opposite wood fibres compared to normal wood, while in the gymnosperm, branching in trachieds was significantly higher in compression wood when compared to opposite wood, demonstrating inconsistent responses between different species. One possible explanation for this low degree of branching could be that at the time at which the samples were fixed, reorientation in induced stems had been completed and therefore branching was no longer required to actively realign microtubule arrays and that the most discordant MTs have already been severed. The higher degree of branching observed in vertical stems could be due to the increased need for more frequent reorientation in response to subtle shifts in dynamic load. This would also explain why no significant differences were observed in branch wood across all species, as load in branch samples would be relatively static.
Other microtubule properties showed either unique, contrasting or no changes between induced and branch reaction wood when compared to normal and opposite wood across species. This lack of consistent responses across states and between divergent woody plants species raises questions regarding if and/or how these structures are involved in reorientation processes and whether different mechanisms exist amongst woody plant species. Such differences between gymnosperm and angiosperm subcellular dynamics are not unprecedented; for example, microtubule organisation is strikingly different between gymnosperm and angiosperm pollen tubes, resulting in different dynamics of intracellular transport and cell wall secretion (Chebli et al 2013).
Immunolabelling techniques employed in the present study are practical tools to observe microtubules; however, due to the nature of the material and the position of xylem cells in a tree’s stem, the use of fixed cells was unavoidable. Microtubules are dynamic polymers, thus images of these arrays in fixed cells are, at best, snapshots of a continuous process of reorganisation at the time of sampling. So far, observation of microtubules in live xylary cells has been achieved, for example, using transdifferentiation of Arabidopsis epidermal cells into tracheary elements (Kubo et al. 2005). While this system has provided valuable insight into secondary cell wall formation, transdifferentiating epidermal cells do not necessarily fully represent xylogenic processes in woody stems across diverse species, especially responses to mechanical stresses or gravitropic change such as reaction wood formation. Together, our results provide the most comprehensive overview of microtubule organisation during reaction wood formation across three woody tree species that show both conserved and species-specific responses to date. Our success in employing an in vitro technique to test the effects of drug treatments on growing xylogenic cells provides useful tools for future cell and molecular biology studies in wood formation within appropriate tissue context. Unravelling exactly how differential cortical microtubule array organisation is triggered during reaction wood formation, how changes in this organisation determine morphogenesis of xylary cells in response to environmental signalling, and whether different tubulin isoforms have different roles in wood formation will be the subject of further investigations.