The phloem transport network is a major evolutionary innovation that enabled plants to dominate terrestrial ecosystems. In the growth apices, the meristems, apical stem cells continuously produce early ‘protophloem’. This is easily observed in Arabidopsis root meristems, in which the differentiation of individual protophloem sieve element precursors into interconnected conducting sieve tubes is laid out in a spatio-temporal gradient. The mature protophloem eventually collapses as the neighboring metaphloem takes over its function further distal from the stem cell niche. Compared with protophloem, metaphloem ontogenesis is poorly characterized, primarily because its visualization is challenging. Here, we describe the improved TetSee protocol to investigate metaphloem development in Arabidopsis root tips in combination with a set of molecular markers. We found that mature metaphloem sieve elements are only observed in the late post-meristematic root, although their specification is initiated as soon as protophloem sieve elements enucleate. Moreover, unlike protophloem sieve elements, metaphloem sieve elements only differentiate once they have fully elongated. Finally, our results suggest that metaphloem differentiation is not directly controlled by protophloem-derived cues but rather follows a distinct, robust developmental trajectory.
The evolution of vascular tissues enabled plants to conquer land because it allowed the separation of the sites of photosynthesis from the sites of nutrient and water acquisition (Lucas et al., 2013). In extant angiosperms, the xylem vessels form hollow tubes to transport water and inorganic ions from the root system to the shoot system. This transport is mainly driven by the water potential differential between the soil and the atmosphere, and therefore by purely physical forces (Endo et al., 2019; Pratt and Jacobsen, 2017). Closely associated with the xylem is the phloem, which is composed of interconnected sieve elements that form the conducting sieve tubes and their neighboring companion cells. Unlike xylem vessels, sieve elements are not dead but, during their differentiation process, they drastically alter their cellular makeup to optimize the transport flow. Most noticeable, they lose their nucleus and vacuole. Thus, sieve elements depend on the neighboring companion cells for the maintenance of their transport functions. Phloem sieve tubes mediate the long distance bulk transport of phloem sap, a viscous mix of sugars, metabolites as well as systemic signaling molecules, from source to sink organs, for example from mature photosynthesizing leaves to roots (Lopez-Salmeron et al., 2019). This transport is driven by a differential in osmotic pressure, which builds up through the controlled loading of osmotic sugars in the source tissue phloem and their unloading in the sink tissue phloem (Knoblauch et al., 2016; Zhang and Turgeon, 2018). The growth apices of plants, the meristems, are terminal sinks, the activity of which is sustained by phloem sap delivered through the early, so-called protophloem. In root meristems, protophloem is produced by apical stem cells that reside adjacent to the quiescent center (QC); it matures while neighboring tissues still divide or undergo expansion growth (Esau, 1977; Lopez-Salmeron et al., 2019). Eventually, its sieve elements become non-functional and are completely obliterated as protophloem is replaced by emerging metaphloem. Although the metaphloem sieve elements share a common precursor with protophloem sieve elements (Bonke et al., 2003; Rodriguez-Villalon et al., 2014), the metaphloem only matures after the expansion growth of the surrounding tissues is completed (Esau, 1977). Metaphloem is then retained as the main conducting phloem, although it can later be replaced by secondary phloem in species that undergo secondary growth.
Non-invasive investigation of phloem development is challenging, first because sieve elements are thin and highly anisotropic cells, and second because the phloem is buried deep inside plant organs. Routine observation of protophloem by confocal microscopy is, however, possible in the root tip of Arabidopsis thaliana, where its development is laid out in a spatio-temporal gradient of ∼20 cells from stem cell daughter to mature sieve element (Furuta et al., 2014; Rodriguez-Villalon et al., 2014). Arabidopsis root tips produce two protophloem strands, which are arranged opposite each other inside the stele, flanking an axis of xylem cells (Fig. 1A). The last two decades have seen tremendous advances in our understanding of protophloem ontogeny. Through its dissection by genetic approaches, numerous protophloem-specific mutants and molecular markers have become available. These studies underline the essential character of root protophloem, the absence or disturbed development of which apparently has grave, systemic consequences on root meristem growth and maintenance (Anne and Hardtke, 2017; Bonke et al., 2003; Rodriguez-Villalon et al., 2014). Whether the defects in the protophloem of pertinent mutants also extend to metaphloem remains largely unknown, mainly because of the difficulty in visualizing metaphloem development and a paucity of specific molecular markers for non-invasive investigation. Here, we set out to mend this gap by developing a toolbox for the analysis of metaphloem development.
RESULTS AND DISCUSSION
An optimized protocol for metaphloem visualization by confocal microscopy
Imaging of the Arabidopsis root tip by confocal microscopy techniques is routine but can be tricky, depending on the tissue targeted for investigation. In particular, this applies to the cells inside the stele, which are small in diameter compared with the surrounding ground tissue or epidermis (Fig. 1A). For example, whereas the diameter of cortex cells reaches ∼25 to ∼50 µm, protophloem cells are a mere ∼5 µm across and therefore roughly 20 times smaller in their horizontal cross-section profile (Fig. 1A). Together, the vascular tissues inside the stele occupy merely ∼10% of the area in a root meristem cross section, although they represent ∼40% of cell files. Despite progress in staining and fixation techniques, visualization of these cells can sometimes be challenging. For example, although developing protophloem sieve elements (PPSEs) can be readily identified because of their early differentiation and associated cell wall build up (Truernit, 2014), mature enucleated PPSEs are difficult to observe. Initially, PPSEs elongate from the ∼20 µm typical for dividing cells to an intermediate stage of ∼50 µm, during which the principal differentiation steps occur. Once they are enucleated, they still elongate rapidly to about twice their length as they become the conductive unloading terminus of the PPSE cell file (Ross-Elliott et al., 2017). It is likely their high anisotropy in combination with a still elongating, soft cell wall that is responsible for the compression of maturing PPSEs by neighboring tissues once they lose their elevated turgor during fixation. Elongating cells possess relatively soft primary cell walls to facilitate directional expansion. Stabilizing secondary cell walls are only deposited during the final stages of differentiation, when the cells have reached their final size and adapt to their future roles. The phenomenon of cell shrinkage or collapse upon fixation is generally observed once all tissues have started to elongate further distal, in the generic cell elongation zone of the root meristem (Fig. S1A). The developing metaphloem sieve elements (MPSEs) are thus particularly affected, rendering their observation difficult with existing standard protocols, such as chloral hydrate clearing (McBryde, 1936) or mPS-PI staining (Truernit et al., 2008). By contrast, the recently developed ClearSee (Kurihara et al., 2015) and TDE (2′2-thiodiethanol) clearing (Musielak et al., 2016) protocols not only preserved the structure of this delicate area (Fig. S1B) but also the fluorescence of reporter proteins.
Starting from these recent advances, we sought to develop a protocol that would leave the elongation area intact and permit routine observation of MPSEs. Through a test series with various combinations and concentrations of described detergents, clearing agents and fixation steps (Kurihara et al., 2015; Musielak et al., 2016; Ursache et al., 2018), we established an optimized procedure that maintained the integrity of the root elongation zone and allowed us to observe the progressive development of MPSE cell files (the ‘TetSee’ protocol, see Materials and Methods) (Fig. S1C). Although the ClearSee protocol delivered very good results for the meristematic regions, we observed shrinkage and deformation of cells in the elongation zone, which was particularly noticeable in 3D image stacks (Fig. S1D,E). The elongating cells were specifically difficult to image because they were susceptible to shrinking and deformation as the stabilizing turgor pressure was removed during fixation and clearing. The optimized TetSee protocol reduced the observed deformations and allowed for a more accurate 3D reconstruction of this area (Fig. S1E). Starting from the second formative division in the phloem lineage, the division that gives rise to the PPSE and MPSE cell files (Bonke et al., 2003; Rodriguez-Villalon et al., 2014), we could thus follow MPSE files across overlapping 3D renderings of serial confocal microscopy images (Fig. 1B). The morphologically visible onset of MPSE differentiation, as judged by intensified Calcofluor White (CCFW) cell wall staining, was on average observed as far as ∼1400 µm from the QC. This was substantially later than the onset of morphological differentiation of PPSEs (∼120 µm from the QC), trichoblasts (∼620 µm from the QC) or protoxylem (∼680 µm from the QC) (Fig. 1C). Thus, with the exception of the metaxylem (which differentiated around the same time or slightly later), MPSEs only differentiated visibly once all other tissues had already matured.
Metaphloem sieve elements differentiate after they have reached their final cell size
Interestingly, whereas cell elongation and differentiation are tightly linked in PPSEs (Furuta et al., 2014; Rodriguez-Villalon et al., 2014), MPSEs elongated to roughly their final size before any cell wall build up became apparent (Fig. 1B). Observation of other cellular rearrangements indicative of MPSE differentiation, notably enucleation, proved to be difficult because of the high anisotropy of MPSEs and against the background from neighboring tissues, for example when nucleic acid dyes such as DAPI were used. However, our morphology-based observations were corroborated by analyses of a generic molecular marker of cellular differentiation in Arabidopsis, the MINIYO (IYO) protein (Sanmartin et al., 2011). The expression of IYO, as well as other markers, was assessed in a similar manner as metaphloem development in wild-type roots, by recording image stacks that covered the vasculature from the meristem up to observable differentiated metaphloem. In all subsequent figures, reporter expression is shown in 3D renderings of these image stacks, which were optimized for the visibility of sieve element cell files and marker gene expression. The original image stacks covered the entire vasculature of the observed root sections and thus allowed the clear identification of the different cell types from the cellular context.
Constitutively expressed IYO-GFP fusion protein is barely visible in the cytosol but accumulates in the nucleus once cells differentiate. In the root tip, IYO-GFP was therefore clearly visible in the quickly differentiating distal root tissues, the columella and lateral root cap (Fig. 2A). Among the proximal root tissues, protophloem is the first to differentiate and, consistently, nuclear IYO-GFP accumulation became first apparent in differentiating PPSEs (Fig. 2A) (Sanmartin et al., 2011). Interestingly, they were followed by their companion cells with some delay, suggesting that PPSE companion cells only differentiate once PPSEs are fully elongated and functional (Fig. 2A). In the stele, developing protoxylem displayed nuclear IYO-GFP next (Fig. 2B), followed, with some delay, by MPSE cell files (Fig. 2C). In fact, nuclear IYO-GFP accumulation was only observed in developing MPSEs after developing protoxylem vessels had already completed their secondary wall build up and after they had themselves fully elongated (Fig. 2D). In summary, both our morphological and molecular analyses suggest that, unlike in PPSEs, cell elongation and terminal differentiation do not coincide in MPSEs.
A set of new molecular markers for the investigation of metaphloem development
Although nuclear IYO accumulation is a very useful generic indicator of the onset of cellular differentiation (Sanmartin et al., 2011), it is not a marker for cell specification. We therefore sought to identify tissue-specific molecular markers that would allow us to trace the incipient beginnings of MPSE development. To this end, we mined the literature for genes that are specifically expressed in mature phloem in other contexts (Anstead et al., 2012; Bonke et al., 2003; Cayla et al., 2015; Khan et al., 2007; Sankar et al., 2014) and chose seven genes for further investigation. Moreover, we intersected existing phloem-related gene expression datasets (Brady et al., 2007; Clark et al., 2019; Kondo et al., 2016; Zhao et al., 2005) to identify a set of 14 additional metaphloem marker candidates. For some of them, existing reporter plasmids could be obtained, but for the majority we cloned promoter constructs that drive the expression of a nuclear localized fluorescent reporter (NLS-CITRINE). After their transformation into Col-0 wild-type plants, eight out of the 21 reporters showed activity in developing root phloem: the previously described SISTER OF ALTERED PHLOEM DEVELOPMENT (SAPL) (Ross-Elliott et al., 2017), EARLY NODULIN-LIKE 9 (ENODL9) (Khan et al., 2007), SIEVE ELEMENT OCCLUSION-RELATED 2 (SEOR2) (Anstead et al., 2012) and SECONDARY WALL-ASSOCIATED NAC DOMAIN PROTEIN 2 (SND2) (Kim et al., 2020) reporters (Fig. 3A-D); and the new reporters DESIGUAL 2 (DEAL2) (Wilson-Sánchez et al., 2018), SIEVE ELEMENT MARKER 1 (SEMA1; AT2G35585), SEMA2 (AT1G61760) and SEMA3 (AT3G26350) (Fig. 4A-D).
However, none of the reporters was exclusively active in the (incipient) metaphloem (Fig. S2A), rather, all markers were also expressed in the late-developing protophloem (Fig. S2B). Among them, the SAPL expression was particular, because although it was expressed in late-differentiating PPSEs similar to the other markers, thereafter it was highly specific for companion cells, both in the proto- and metaphloem, and not detected in developing MPSEs (Fig. 3A). Notably, SAPL was continuously expressed from the early coincidence with PPSE differentiation onward beyond differentiated MPSEs and was not observed in any other cell file. This suggests that the four companion cell files subsequently serve both PPSE and MPSE maintenance. The other markers were expressed in PPSEs as well as MPSEs, with varying levels of specificity. All of them were expressed in developing PPSEs, after the onset of cell wall build up and coincident with the partial elongation that occurs before enucleation. ENODL9, SEOR2 and SEMA1 were most specific for developing sieve elements (Figs 3C,D and 4B). However, whereas SEOR2 and SEMA1 expression gradually ceased upon PPSE differentiation and only became active again later, ENODL9 expression switched to the incipient MPSE file earlier and stayed on until MPSE differentiation terminated (Fig. 3D). Moreover, SEOR2 expression reappeared earlier than SEMA1 expression (Figs 3C and 4B). The other reporters also displayed some marked expression outside of PPSEs/MPSEs. The SND2 reporter was strongly expressed in late-developing MPSEs; however, it was also observed in developing metaxylem (Fig. 3B). DEAL2, SEMA2 and SEMA3 all switched expression to the cell files surrounding PPSEs after enucleation (Fig. 4A,C,D). In the metaphloem, DEAL2 was expressed in MPSEs but also in the directly neighboring cell files, likely the companion cells (Fig. 4A). A similar pattern was observed for SEMA3 (Fig. 4D), whereas SEMA2 appeared to be specific for MPSEs (Fig. 4C). In summary, we were able to identify a set of reporters for metaphloem development that mark different stages as well as cell types (Fig. 4E). Their investigation confirmed that, unlike what has been reported for PPSEs, cell elongation and differentiation are uncoupled in MPSEs, and also showed that both sieve element types are associated with the same companion cell files.
Metaphloem development is not affected by CLE45 treatment
The continuous expression of ENODL9 in the MPSE cell files as soon as PPSEs enucleate also suggested that MPSE specification starts as soon as PPSE development is finished. This could mean that premature MPSE differentiation is prevented by lateral inhibition through cues derived from developing PPSEs. One such candidate signal is secreted CLAVATA3/EMBRYO SURROUNDING REGION-RELATED (CLE) signaling peptides, because low concentrations of certain synthetic CLE peptides suppress PPSE development when applied to roots (Depuydt et al., 2013; Hazak et al., 2017; Ito et al., 2006; Kinoshita et al., 2007; Rodriguez-Villalon et al., 2014), as does dosage increase of CLE45 (Czyzewicz et al., 2015b; Rodriguez-Villalon et al., 2014). Interestingly, CLE45 as well as CLE26 and CLE25 are specifically expressed in developing PPSEs (Czyzewicz et al., 2015a; Ren et al., 2019; Rodriguez-Villalon et al., 2014, 2015). CLE peptide signaling is, however, apparently not strictly required for protophloem development (Anne et al., 2018; Fukuda and Hardtke, 2020), rather it appears to act as a safeguard mechanism that maintains plasticity of phloem pole cells during their meristematic stage (Gujas et al., 2020).
Upon CLE45 treatment, the expression of both markers tested, SEMA3 and SEOR2, disappeared from the protophloem, consistent with the prohibitive CLE45 effect on PPSE formation (Fig. 5A-D). However, both markers persisted in developing MPSEs (Fig. 5B,D), in line with the observation that their differentiation appeared to be unaffected. Notably, this observation also confirmed once more that MPSE specification is position- rather than lineage-dependent, because the PPSE and MPSE cell files arise from the same stem cell daughter through a periclinal division that is suppressed by CLE45 application (Rodriguez-Villalon et al., 2014, 2015). Moreover, the absence of root growth defects in cle25 mutants (Ren et al., 2019), as well as in receptor mutants that are fully insensitive against all three CLE peptides (Anne et al., 2018), corroborates the conclusion that PPSE-derived CLE peptides do not impinge on MPSE development under normal circumstances. In summary, CLE45 peptide treatment efficiently suppressed PPSE formation, but did not interfere with MPSE development.
MPSE development follows a robust developmental trajectory
In the protophloem, CLE45 signaling is quantitatively antagonized by the vascular plant-specific OCTOPUS (OPS) gene. OPS is thus a positive regulator of PPSE differentiation that is expressed from early on in the protophloem and insulates developing PPSEs against the effects of autocrine CLE45 signaling (Breda et al., 2017, 2019). In ops loss-of-function mutants, developing PPSEs frequently fail to differentiate (including failure to build up cell wall and thereby appearing as so-called gap cells), which causes discontinuities in the protophloem strands and disturbs the transport of phloem sap into the meristem (Anne and Hardtke, 2017; Rodriguez-Villalon et al., 2014; Truernit et al., 2012). OPS is also weakly expressed in the incipient MPSE cell file, against a background of low, ubiquitous expression of its homolog OPS-LIKE 2 (OPL2) that increases in developing metaphloem (Ruiz Sola et al., 2017). Whereas opl2 single mutants did not display apparent phenotypes, except a more variable root growth vigor (Fig. S3A), the ops opl2 double mutant is the only described genotype with MPSE defects so far (Fig. S3B) (Ruiz Sola et al., 2017), apart from mutants that lack protophloem and metaphloem altogether (e.g. Bonke et al., 2003; Scheres et al., 1995). Compared with ops single mutants, root growth vigor was further diminished in ops opl2 double mutants (Fig. S3A,C) and they also displayed aggravated PPSE differentiation defects (Fig. S3B,D) (Ruiz Sola et al., 2017). The latter were more severe than evident from simple gap cell presence-absence counts, because ops opl2 double mutants often had only one distinguishable PPSE strand. To better understand how MPSE and PPSE differentiation is affected in ops single and ops opl2 double mutants, we introduced some of our reporter genes into these backgrounds.
In ops single mutants, the SEMA3 reporter was expressed at the later stages of PPSE differentiation, as in wild type, but absent in developing PPSEs that failed to differentiate (Fig. S4A), underlining their different cellular identity. In developing MPSEs, SEMA3 expression appeared to be unaffected (Fig. S4B,C). By contrast, SAPL reporter activity was still observed in gap cells (Fig. S4D), which corroborates earlier observations and is in line with the recent proposal that they adopt companion cell identity (Gujas et al., 2020) as well as the strong continuous companion cell-specific SAPL expression after PPSE differentiation. Again, SAPL expression appeared to be unaffected in the developing metaphloem region of ops mutants (Fig. S4E). Together, these findings reiterate that the defects in ops mutants are protophloem-specific (Ruiz Sola et al., 2017; Truernit et al., 2012). Interestingly, SAPL expression could still be detected in protophloem gap cells later on (Fig. S4F), indicating that PPSE cells that fail to differentiate properly in the protophloem differentiation window fail to catch up.
In ops opl2 double mutants, the SEOR2, DEAL2, SEMA2 and SEMA3 markers displayed normal expression, except their apparent absence in gap cells (Fig. 6A-D). Despite the described MPSE differentiation defects (Ruiz Sola et al., 2017), which we could also observe in optical cross-sections (Fig. S3B), our markers were essentially continuously expressed in developing MPSEs of ops opl2 mutants (Fig. 6A-D). Thus, we could not detect corresponding ‘metaphloem gap cells’, possibly because differentiating MPSEs are quite long (200-300 µm) and because surveying extended stretches of MPSEs was difficult. Nevertheless, although ops opl2 double mutants typically displayed marker expression in both PPSE and MPSE strands (Fig. 7A), this pattern also often deviated from wild type. Upon closer inspection, this could be attributed to the reappearance of reporter expression in undifferentiated protophloem cell files long after the zone of normal PPSE differentiation, clearly visible from elongated PPSEs that expressed the respective marker (Fig. 7B). Thus, the observation that one of the two protophloem poles in ops opl2 mutants was frequently absent (Fig. S3B) (Ruiz Sola et al., 2017) could also reflect a strongly delayed differentiation of one PPSE strand. In extremis, the delay was such that it overlapped with reporter expression in the neighboring MPSE cell files (Fig. 7C). We had not observed such atypical differentiation in ops single mutants. This not only indicates that PPSE differentiation can be substantially delayed in ops opl2 mutants, but also that the onset of MPSE differentiation is largely independent of such delays. Corroborating the independent trajectory of MPSE differentiation, in ops opl2 roots, in which only one PPSE cell file showed differentiating cells (any gap cells notwithstanding) and marker expression was also absent from the failed PPSE cell file later on, marker expression in both MPSE cell files appeared to be normal (Fig. 7D) and could typically be observed shortly after protoxylem cells with secondary cell walls were visible, as in wild type. Together with the observed activity of our markers in CLE45-treated roots, our analyses therefore suggest that MPSE development follows a robust trajectory that is largely independent from PPSE development.
Finally, it is noteworthy that OPS action is exquisitely dosage-sensitive (Breda et al., 2017, 2019) and promotes PPSE differentiation by quantitatively antagonizing CLE45 signaling via the receptor kinase BARELY ANY MERISTEM 3 (Breda et al., 2017; Breda et al., 2019; Fukuda and Hardtke, 2020). Moreover, an excess of ectopic OPS activity leads to premature differentiation across root tissues (Breda et al., 2019). Thus, our results are consistent with the notion that the differential expression levels of OPS family proteins such as OPS and OPL2 along the gradient of developing PPSE versus MPSE cell files contributes to the correct spatio-temporal separation of their differentiation.
A toolbox for the investigation of metaphloem development
In summary, our study extends our toolbox for the investigation of sieve element development in the Arabidopsis root, with a special focus on the so far poorly described differentiation of the metaphloem, and provides first forays into its genetic control. Our results highlight commonalities between PPSE and MPSE development, but also suggest that metaphloem development follows a robust trajectory that is not directly influenced by adjacent or preceding PPSE development under normal circumstances. Combined with state-of-the-art technical advances, such as single cell RNA-sequencing or tissue-specific gene knockout (Smetana et al., 2019; Wendrich et al., 2020), our observations should enable more targeted future approaches to dissect metaphloem development and discover its unique features.
MATERIALS AND METHODS
Plant materials and growth conditions
Seeds were surface sterilized using 3% sodium hypochlorite, sown onto half strength Murashige & Skoog (MS) agar medium (0.9% agarose) supplemented with 0.3% sucrose and stratified for 3 days at 4°C. Plants were grown under continuous white light (intensity ∼120 µE) at 22°C. All mutants and marker lines were in the Arabidopsis thaliana Columbia-0 (Col-0) wild-type background. The ops and ops opl2 mutant lines have been previously described (Ruiz Sola et al., 2017; Truernit et al., 2012). CLE45 peptide treatments were performed as previously described (Anne et al., 2018).
Database mining and selection of sieve element marker candidates
For the selection of the sieve element (SE) marker candidates, expression of 576 genes enriched in cells expressing the S32 phloem marker (AT2G18380) was analyzed along the root, in cells expressing SUC2 (AT1G22710) (Brady et al., 2007), in cells expressing CVP2 (AT1G05470) (Clark et al., 2019), in a general root and seedling gene expression dataset (Gan et al., 2011), and in the ‘VISUAL’ phloem and xylem datasets (Kondo et al., 2016). Twenty candidate genes that: (1) showed expression in the phloem poles, (2) showed increased expression further away from the meristem, (3) showed relatively higher expression in the root, and (4) appeared in the VISUAL phloem datasets were tested as SE markers.
For the construction of SE markers, respective promoter fragments of 1500-2500 bp were amplified from genomic Col-0 DNA using suitable oligonucleotides with overhangs (attB1/2 extensions for ENODL9, SEOR2, SND2, and attB4/1r extensions for DEAL2 and SEMA1-3) for subsequent Gateway™ cloning (see Table S1). The manufacturer's solutions (Thermo Fisher Scientific article numbers 11791020 and 11789020) and protocols were used for all cloning reactions. Amplified fragments were cloned into suitable entry vectors and the ENODL9, SEOR2 and SND2 promoters were transferred into the pMDC205 destination vector in front of a GFP reporter with an endoplasmic reticulum (ER) retention signal (Curtis and Grossniklaus, 2003). The DEAL2 and SEMA1/2/3 promoters were recombined together with NLS-CITRINE in a multisite gateway reaction into the pK7m24 vector backbone. Flowering Col-0 plants were transformed using the floral dip method and transformants were selected either on ½ MS media containing 25 mg/ml hygromycin B or 25 mg/ml kanamycin following a fast selection procedure (Harrison et al., 2006).
Tissue fixation and clearing (TetSee protocol)
For microscopy, 7-day-old seedlings were fixed in a solution of 4% paraformaldehyde in PBS buffer and transferred into a vacuum of 25-30 mmHg/Torr for 15-30 min. Subsequently, seedlings were washed three times in PBS for 5 min. For clearing of the samples, different protocols were used and assessed for the quality of the tissue preservation along the root. Although standard protocols like chloral hydrate clearing (McBryde, 1936) or mPS-PI staining (Truernit et al., 2008) caused the shrinking of the cells in the early elongation zone, the recently developed ClearSee (Kurihara et al., 2015) and TDE clearing (Musielak et al., 2016) protocols preserved the structure of this delicate area as well as the fluorescence of reporter proteins. The two protocols were further optimized and combined into the ‘TetSee’ (2′2-Thiodiethanol-ClearSee) protocol. Briefly, the washed seedlings were transferred into TetSeeX solution [15% Na-deoxycholate, 25% urea, 10% glycerol, 5% TDE (Merck, 166782), 1% Triton X-100] and kept for 3 days at 4°C with daily changes of the TetSeeX solution. For microscopy, the TetSeeX solution was removed and replaced by TetSee solution (the TetSeeX solution without Triton X-100) containing 0.25 mg/ml CCFW (Sigma-Aldrich, F3543). Seedlings were incubated in the CCFW staining solution for 6 h or overnight, washed once in TetSee solution and then transferred onto microscopy slides with TetSee solution as mounting medium.
For morphological assessment of phloem differentiation, roots were prepared as described above and the CCFW-stained cell walls were imaged using a Zeiss LSM880 confocal microscope with a 40× objective. A 405 nm laser was used for CCFW excitation, and the cell wall signal was recorded in a range from 450-480 nm. For imaging of the SE marker lines, GFP or CITRINE were sequentially excited with 488 nm and their emission recorded from 500-560 nm. Tile scans and z-scans were combined in order to obtain continuous images of the vasculature from the root meristem to the differentiated metaphloem. In addition, a Nikon spinning disc CSU-W1 confocal microscope with a 40× objective was used to record images of the CLE45-treated and the ops opl2 SE marker lines. Analysis of the images and generation of 3D renderings from the z-stacks were performed using the GNU icy software.
We would like to thank Dr Y. Helariutta for a gift of the pSAPL::erGFP reporter plasmid, and Drs J. Sanchez-Serrano and E. Rojo for a gift of the p35S::IYO-GFP plasmid and transgenic line.
Conceptualization: M.G., C.S.H.; Methodology: M.G.; Validation: M.G., C.S.H.; Investigation: M.G.; Data curation: M.G.; Writing - original draft: M.G., C.S.H.; Writing - review & editing: M.G., C.S.H.; Supervision: C.S.H.; Project administration: C.S.H.; Funding acquisition: M.G., C.S.H.
This work was funded by Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung grant 310030B_185379 (awarded to C.S.H.) and the Deutsche Forschungsgemeinschaft post-doctoral fellowship GR 5009/1-1 (awarded to M.G.).
The authors declare no competing or financial interests.