Endocytic trafficking promotes vacuolar enlargements for fast cell expansion rates in plants

  1. Kai Dünser
  2. Maria Schöller
  3. Ann-Kathrin Rößling
  4. Christian Löfke
  5. Nannan Xiao
  6. Barbora Pařízková
  7. Stanislav Melnik
  8. Marta Rodriguez-Franco
  9. Eva Stöger
  10. Ondřej Novák
  11. Jürgen Kleine-Vehn  Is a corresponding author
  1. Molecular Plant Physiology (MoPP), Faculty of Biology, University of Freiburg, Germany
  2. Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Austria
  3. Center for Integrative Biological Signalling Studies (CIBSS), University of Freiburg, Germany
  4. Laboratory of Growth Regulators, Institute of Experimental Botany of the Czech Academy of Sciences and Faculty of Science of Palacký University, Czech Republic
  5. Cell Biology, Faculty of Biology, University of Freiburg, Germany

Abstract

The vacuole has a space-filling function, allowing a particularly rapid plant cell expansion with very little increase in cytosolic content (Löfke et al., 2015; Scheuring et al., 2016; Dünser et al., 2019). Despite its importance for cell size determination in plants, very little is known about the mechanisms that define vacuolar size. Here, we show that the cellular and vacuolar size expansions are coordinated. By developing a pharmacological tool, we enabled the investigation of membrane delivery to the vacuole during cellular expansion. Our data reveal that endocytic membrane sorting from the plasma membrane to the vacuole is enhanced in the course of rapid root cell expansion. While this ‘compromise’ mechanism may theoretically at first decelerate cell surface enlargements, it fuels vacuolar expansion and, thereby, ensures the coordinated augmentation of vacuolar occupancy in dynamically expanding plant cells.

Editor's evaluation

Plant cells can grow to extraordinarily large volumes; Arabidopsis root cells, for example, can expand beyond 50um long. Vacuole expansion is correlated with cell elongation, presumably to "fill up" the volume of the cell without requiring a tremendous volume of cytoplasm. Here, the authors carefully characterize an exciting new small molecule inhibitor of endocytic trafficking to the vacuole. This new tool will be valuable to researchers studying endocytic trafficking and vacuole biogenesis in plants.

https://doi.org/10.7554/eLife.75945.sa0

Introduction

Animals and plants take many shapes and come in great diversity of sizes (Marshall et al., 2012). During cellular expansion, the surface area to intracellular volume ratio gets smaller and if cells grow beyond a critical limit, the surface of the plasma membrane may cease to accommodate cellular needs. Cells typically remain, hence, relatively small or eventually induce membrane furcation to enlarge the cellular surface. Compared to animals, plant cells can dramatically increase their size without the apparent need for surface furcation. The vacuole, occupying up to 90% of a cellular volume, is the biggest organelle in plants and its size correlates with and determines the cell size (Owens and Poole, 1979; Berger et al., 1998; Löfke et al., 2015; Scheuring et al., 2016). The size of the vacuole increases during cell expansion and vacuolar morphology is a reliable intracellular read-out for cellular expansion (Dünser et al., 2019). The dynamic regulation of vacuolar size allows plant cells to expand with little increase of the cytosol (Dünser et al., 2019), maintaining a favorable cell surface to cytosol ratio during growth. Despite its anticipated importance for fast cellular expansion, very little is known about how cells dynamically scale and coordinate its vacuolar size (Dünser and Kleine-Vehn, 2015; Krüger and Schumacher, 2018). Vacuolar soluble N-ethylmaleimide-sensitive-factor attachment receptors (SNAREs) reliant membrane remodeling as well as actin/myosin-dependent vacuolar constrictions define the auxin-dependent vacuolar occupancy of the cell (Löfke et al., 2015; Scheuring et al., 2016). The vacuolar SNARE complex, consisting of R-SNARE VAMP711, Qa-SNARE SYP22, Qb-SNARE VTI11, as well as Qc-SNAREs SYP51 and SYP52, regulates hetero- and homotypic membrane fusion at the tonoplast (Fujiwara et al., 2014). Despite our molecular knowledge on vesicle trafficking toward the vacuole (Krüger and Schumacher, 2018; Kang et al., 2022), very little is known about the dynamics of membrane delivery to the vacuole and its implication for cellular expansion.

Results

In this project, we hypothesized that the interference with vacuolar SNAREs could aid us to visualize the rate and importance of membrane delivery to the vacuole. Mutations in the vacuolar SNARE subunit VTI11 leads to strongly affected, roundish, and partially fragmented vacuoles (Yano et al., 2003; Zheng et al., 2014; Löfke et al., 2015). Despite the strongly affected vacuolar morphology, the root length, as well as the size of fully elongated epidermal cells of untreated vti11 mutants, were hardly distinguishable from wild-type seedlings (Figure 1A-D). This finding seemingly questions the contribution of vacuolar SNARE-dependent membrane fusions at the tonoplast for cellular expansion. Considering that the space-filling function of the vacuole defines cell size determination (Dünser et al., 2019), we next analyzed the contribution of VTI11 for vacuolar occupancy of the cell, using confocal z-stack imaging and 3D renderings of epidermal atrichoblast cell files in the meristematic and elongation zone (Figure 1E). Although vti11 mutant vacuoles are morphologically distinct from wild-type vacuoles (Yano et al., 2003; Zheng et al., 2014; Löfke et al., 2015), their vacuolar occupancy of epidermal root cells was not distinguishable from wild-type (Figure 1F). This finding suggests that redundancy may ensure the space-filling function of the vti11 mutant vacuole and, thereby, likely safeguards cell expansion.

Vacuolar occupancy of the cell defines cell expansion.

(A) Representative images (scale bars: 0.5 cm) and (B) quantification of the main root length of 6-day-old Col-0 (n=36) and vti11 seedlings (n=32). Student’s t-test (ns). (C) Representative images (scale bars: 25 µm) of differentiated atrichoblast cells. Seedlings were stained with propidium iodide (PI) (red) for 30 min in liquid medium prior image acquisition. (D) Boxplots show cell length quantification of 6-day-old Col-0 (n=53) and vti11 (n=59) seedlings. Student’s t-test (ns). (E) 3D reconstructions of PI-stained cell walls (gray) and BCECF-stained vacuoles (magenta) in the early and late meristem and in the early and late elongation zone. Scale bars: 5 µm. (F) Boxplots show vacuolar occupancy of cells in the defined zones (n=7–16). Student’s t-test (ns). Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values. Representative experiments are shown.

VTI11 and VTI12 are able to substitute each other in their respective SNARE complexes with the vti11 vti12 double mutants being embryonic lethal (Surpin et al., 2003). In order to overcome genetic redundancy as well as lethality in the crucial control of vesicle trafficking toward the vacuole, we conducted a small molecule screen with the aim to identify compounds impacting on SNARE-dependent delivery of vesicles to the vacuole. For our primary screen, we germinated seedlings constitutively expressing the vacuolar R-SNARE marker pUBQ10::YFP-VAMP711 in growth medium supplemented with bioactive chemicals from a library of 360 compounds, which likely affect cell expansion in planta (Drakakaki et al., 2011; Figure 2A). We subsequently employed a fluorescence binocular to identify compounds that intensified the YFP-VAMP711 signal (Figure 2B). In a secondary, CLSM (confocal laser scanning microscopy)-based validation screen, we recorded vacuolar morphology changes in the samples treated with the compounds in comparison to solvent (control) treatments (Figure 2C, Figure 2—figure supplement 1). Thereby, we isolated 12 small molecules (Figure 2—figure supplement 1) that affected YFP-VAMP711 and vacuolar morphology (Figure 2—figure supplement 1, Figure 2—figure supplement 2), which we thus named vacuole affecting compounds (VACs) 1–12.

Figure 2 with 5 supplements see all
Bioactive small molecule screen identifies VAC1 as a vacuole affecting compound.

(A) Schematic depiction of the small molecule screen workflow. pUBQ10::YFP-VAMP711 seeds were germinated in liquid medium containing solvent control dimethyl sulfoxide (DMSO) or 360 small molecules from a library of bioactive compounds. (B) Four-day-old seedlings were then screened for intensified YFP-VAMP711 signal using a fluorescence binocular. Scale bars: 0.5 cm. (C) Representative images of the confocal laser scanning microscopy (CLSM)-based confirmation screen. Cell wall and vacuolar membrane in late meristematic atrichoblast cells were visualized with propidium iodide (PI) (green) and pUBQ10::YFP-VAMP711 (yellow), pSYP22::SYP22-GFP (yellow), or pUBQ10::pHGFP-VTI11 (yellow), respectively. Scale bars: 5 µm. (D–F) Representative transmission electron microscopy (TEM) images of solvent control (DMSO)- (D) and VAC1-treated (E) epidermal root cells. (F) shows a magnified detail of (E). Scale bars: 5 µm (D,E) and 1 µm (F).

Here, we primarily report on VAC1 (N’-[(2,4-dibromo-5-hydroxyphenyl)methylidene]-2-phenylacetohydrazide), which affected vacuolar morphology (Figure 2—figure supplement 1) and induced ectopic accumulation of vacuolar SNARE marker YFP-VAMP711 in aster-like structures adjacent to the main vacuole (Figure 2C). Due to the possibly hydrolyzable acylhydrazone bond in VAC1, we wondered whether VAC1 could be potentially metabolized in planta into some of its sub-structures or other metabolic products. VAC1 can be synthesized from 2,4-dibromo-5-hydroxybenzaldehyde and phenylacetic acid (PAA) hydrazide, but the application of these substances (here referred to as precursors) did not cause any alterations to vacuolar morphology (Figure 2—figure supplement 3A). We also analyzed the potential VAC1 conversion to PAA, because it is a naturally occurring compound with low auxin activity (Abe et al., 1974). The ultra-high performance liquid chromatography-selected ion recording-mass spectrometry (UHPLC-SIR-MS) analysis however did not detect PAA in VAC1-treated seedlings (Figure 2—figure supplement 3B-D).

We conclude that VAC1 is not detectably converted to PAA under our conditions. On the other hand, in VAC1-treated samples, the UHPLC-UV-MS analysis revealed three additional peaks for potential VAC1 metabolites (M1-M3) with a retention time of 4.37, 5.26, and 5.39 min for M1, M2, and M3, respectively (Figure 2—figure supplement 3E-F). High-resolution mass spectrometry (HRMS) identified these metabolites based on exact mass determination and characteristic distribution of bromine isotopes (Figure 2—figure supplement 3G). Metabolite M1 was predicted to be the hydroxylated form (VAC-OH, Rt 4.37 min, [M-H-]: 424.9129 Da, calculated elemental composition C15H11Br2N2O3) and the other two derivatives relate to glycosylated forms. The detected compound M2 likely represented VAC1-3-O-glucoside (VAC1-Glc, Rt 5.26 min, [M-H-]: 570.9706 Da, C21H21Br2N2O7) and M3 was identified as VAC1-3-O-6-O-acetylglucoside (VAC1-AcGlc, Rt 5.39 min, C23H23Br2N2O8, [M-H-]: 612.9805 Da) (Figure 2—figure supplement 3G). Using UHPLC–UV (λmax = 291 nm for all compounds), we revealed VAC1-AcGlc (56.3% ± 0.8%) as the most abundant metabolite, followed by VAC1-Glc (31.3% ± 0.4%) and VAC1 (12.4% ± 0.6%), while the UV signal of hydroxylated VAC1-OH (<0.1%) was close to the detection limit of the photodiode array detector used (Figure 2—figure supplement 3H). This set of data shows that a substantial amount of VAC1 undergoes glycosylation in planta, which could affect its activity in long-term experiments.

Next, we used structurally related derivatives of VAC1, termed VAC1-analogue (VAC1A), to further address whether the entity of VAC1 is required for its effects on vacuolar morphology (Figure 2—figure supplement 4A). At the concentration and treatment time used for the subcellular effect of VAC1 (10 µM for 2.5 hr), solely VAC1A4, which displays an additional hydroxy group, induced severe vacuolar morphology defects (Figure 2—figure supplement 4B). The cellular effects of VAC1A4 were, however, visually distinct from VAC1. While short-term (2.5 hr) VAC1A4-treated cells appeared viable (based on the intact extracellular propidium iodide [PI] staining) (Figure 2—figure supplement 4B), the main root growth did not resume after wash-out (Figure 2—figure supplement 4C), indicating some degree of toxicity. On the other hand, interference with either the phenyl group or the functional groups of the 2,4-dibromo-5-hydroxybenzaldehyde reduced the in planta activity and/or stability of VAC1 (Figure 2—figure supplement 4B). At higher concentrations (25 and 50 µM), VAC1A2 and VAC1A5 also impacted to some degree on vacuolar morphology, while VAC1A1 and VAC1A3 appeared largely inactive in regard to vacuolar morphology (Figure 2—figure supplement 4B). These results imply that the entity of VAC1 (phenyl- as well as the benzaldehyde rings) contributes to its effect on vacuolar morphology.

Next, we addressed the subcellular effects of VAC1 on the endomembrane system. When compared to YFP-VAMP711, we observed similar VAC1 effects on vacuolar morphology and strong aster-like protein accumulation of pUBQ10::pHGFP-VTI11- and pSYP22::SYP22-GFP- (Figure 2C). This finding suggests that VAC1 affects the distribution of vacuolar SNARE complex components. VAC1 also induced similar ectopic accumulation of vacuolar membrane marker VHA-a3-GFP (Figure 2—figure supplement 5), suggesting that other tonoplast proteins are not excluded from the region of aster-like aggregations and that these structures are part of the tonoplast. To obtain a better understanding of the VAC1-reliant subcellular defects, we subsequently performed electron microscopy. When compared to control treatments, VAC1 induced ectopic extensions of the vacuole, which were reminiscent to the before mentioned aster-like structures (Figure 2D and E). In addition, we observed VAC1-dependent clusters of vesicles in close proximity to the malformed vacuole (Figure 2F), possibly visualizing defective vesicle delivery to the vacuole. To address whether VAC1 disrupts vesicle trafficking to the vacuole, we next employed the styryl dye FM4-64 (Vida and Emr, 1995; Scheuring et al., 2015). The endocytic tracer FM4-64 initially labels the plasma membrane and is rapidly internalized following the endocytic pathway, eventually staining the vacuolar membrane (Ueda et al., 2001; Scheuring et al., 2015). Tonoplast labelling with FM4-64 was observed within 3 hr following the pulse staining in mock-treated seedlings (Figure 3A). Strikingly, staining of the vacuolar membrane was absent in VAC1-treated samples and instead FM4-64 accumulated in roundish structures (VAC1 bodies) (Figure 3A). This set of data demonstrates that VAC1 treatments prevent the fusion of FM4-64-positive endocytic vesicles to the central vacuole.

VAC1 interferes with vesicle trafficking to the vacuole.

(A) Tonoplast and plasma membrane of late meristematic atrichoblast cells were visualized with FM4-64 (magenta) and pUBQ10::NPSN12-YFP (gray), respectively. Six-day-old seedlings were pre-treated with dimethyl sulfoxide (DMSO) or 10 µM VAC1 for 30 min., then pulse-stained with 4 µM FM4-64 for 5 min and subsequently transferred to liquid medium containing DMSO or 10 µM VAC1 for 3 hr. Scale bars: 5 µm. (B) Representative images of 6-day-old seedlings of plasma membrane marker lines pBRI1::BRI1-GFP (yellow), pPGP19::PGP-19-GFP (yellow), and pUBQ10::NPSN12-YFP (yellow). Seedlings were treated with DMSO or 10 µM VAC1 for 2.5 hr in liquid medium. Scale bars: 5 µm. (C) Representative images of 6-day-old seedlings of CLC-GFP (green) and Lifeact-venus (yellow) marker lines. Seedlings were treated with DMSO or 10 µM VAC1 for 2.5 hr in liquid medium. Scale bars: 5 µm. (D) Representative images of 6-day-old WAVE18Y (pUBQ10::GOT1-YFP) (yellow), pVHAa1::VHAa1-GFP (yellow), and WAVE7Y (pUBQ10::RABF2a-YFP) (yellow) treated with DMSO or 10 µM VAC1 for 2.5 hr in liquid medium. Cell walls were counterstained with propidium iodide (PI). Scale bars: 5 µm. (E) Six-day-old p35S::SYP21-GFP (magenta) seedlings were pulse-stained with 4 µM FM4-64 (gray) for 5 min and subsequently pre-treated with 25 µM BFA for 30 min before adding 10 µM VAC1 or solvent control for another 60 min. The treatments were conducted in liquid medium. Scale bars: 5 µm. (F) Boxplots show mean gray values of FM4-64 fluorescence signal in VAC1 bodies in VAC1 (n=24) or VAC1 + BFA (n=24) treatments. Student’s t-test (****p<0.0001). Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values. Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values.

Next, we inspected plasma membrane resident proteins BRI1-GFP, PGP19-GFP, and NPSN12-GFP, expecting a fraction of these proteins transiting to the vacuole for lytic degradation (Kleine-Vehn et al., 2008). When compared to the ectopic FM4-64 accumulations, the examined lines showed comparably weak, but clearly detectable VAC1-induced aggregation of the fluorescently tagged proteins, which were absent in the solvent control treatments (Figure 3B). This finding corroborates our assumption that VAC1 interferes with vesicle trafficking to the vacuole. On the other hand, VAC1 did not visibly affect clathrin light chain marker CLC-GFP and F-actin marker Lifeact-venus (Figure 3C). Similarly, early and late endosome marker lines, such as GOT1-YFP (WAVE18Y), VHA-a1-GFP, and RABF2A-YFP (RHA1), were not disrupted upon VAC1 administration (Figure 3D). This set of data suggests that early vesicle sorting toward the vacuole is likely not disturbed, but stalls in close proximity to the vacuole.

To further assess VAC1-sensitive vesicle trafficking, we initially used brefeldin A (BFA), which is a specific inhibitor of ARF-GEFs (ADP-ribosylation factor – guanine-nucleotide exchange factors) and blocks endocytic trafficking toward the vacuole at the level of the trans-Golgi network (TGN) (Kleine-Vehn et al., 2008). The VAC1-induced accumulation of vacuolar SNARE SYP21-GFP surrounded the FM4-64 accumulation in VAC1 bodies, which was significantly reduced in BFA co-treatments (Figure 3E and F). We, hence, conclude that VAC1 affects vacuolar vesicle trafficking downstream of the TGN. To further decipher the VAC1 effect downstream of the TGN, we next addressed its dependency on the CLASS C CORE VACUOLE/ENDOSOME TETHERING (CORVET) and homotypic fusion and protein sorting (HOPS) machineries (Takemoto et al., 2018), because these membrane tethering complexes supply the vacuolar SNAREs with heterotypic (endosomal derived) and homotypic (vacuole derived) membranes, respectively. The dexamethasone (DEX)-inducible amiRNA lines that target either the CORVET-specific subunit VPS3 or the HOPS-specific subunit VPS39 enable the selective repression of the CORVET and HOPS complex, respectively (Takemoto et al., 2018). The transcriptional repression of CORVET or HOPS inhibits root growth (Figure 4A and B and Figure 4—figure supplement 1A-C). In agreement, seedlings germinated on VAC1 containing media also displayed reduced main root growth (Figure 4A, B, F, G), suggesting that vesicle trafficking to the vacuole contributes to root growth. Homotypic membrane fusion unlikely contributes to the VAC1 effect, because VPS39 depleted roots remained fully sensitive to VAC1 (Figure 4—figure supplement 1A-C). In contrast, the repression of CORVET function induced a partially resistant root growth to VAC1 (Figure 4A–C), suggesting that defects in heterotypic vesicle trafficking toward the vacuole contribute to the VAC1-dependent interference with root growth. To visualize vacuolar morphology after CORVET repression, we used the pUBQ10::YFP-VAMP711 marker crossed to the DEX-inducible VPS3 amiRNA line (Takemoto et al., 2018). The conditional knockdown of VPS3 had no visible effect on the distribution of YFP-VAMP711 and the VPS3-deprived cells remained sensitive to VAC1-induced aberrations in VAMP711 localization (Figure 4D).

Figure 4 with 1 supplement see all
VAC1 specifically interferes with vacuolar soluble N-ethylmaleimide-sensitive-factor attachment receptor (SNARE)-dependent vesicle fusion to the tonoplast.

(A) Representative images (scale bars: 0.5 cm) and (B) quantification of main root length of 7-day-old amiR Vps3 seedlings germinated on solvent control medium (dimethyl sulfoxide [DMSO], n=27), 10 µM VAC1 (n=25), 30 µM dexamethasone (DEX, n=19) or 10 µM VAC1 and 30 µM DEX (n=17). Two-way ANOVA with Sidaks’s multiple comparisons test (DMSO to VAC1: ****p<0.0001, DEX to DEX + VAC1: p=0.3247). (C) Boxplots show relative growth of VAC1 samples and DEX + VAC1 samples. Statistics is based on two-way ANOVA and multiple comparison analysis (DMSO to VAC1: ****p<0.0001, DEX to DEX + VAC1: p=0.0660). (D) Representative images of late meristematic atrichoblast cells in amiR Vps3 x pUBQ10::YFP-VAMP711 seedlings. Seedlings were grown for 4 days, then transferred to plates containing solvent control (DMSO) or 30 µM DEX for another 3 days and subsequently treated for 2.5 hr with DMSO, 30 µM DEX, 10 µM VAC1 or VAC1 + DEX, respectively, in liquid medium. Scale bars: 5 µm. (E) Representative images of late meristematic atrichoblast cells of pUBQ10::YFP-VAMP711 in wild-type and vti11 background. Propidium iodide (PI) (green) and pUBQ10::YFP-VAMP711 (yellow) depict cell wall and vacuolar membrane, respectively. Six-day-old seedlings were treated with solvent control (DMSO) or 10 µM VAC1 for 2.5 hr in liquid medium. Scale bars: 5 µm. (F) Representative images (scale bars: 0.5 cm) and (G) boxplots showing main root length of 6-day-old Col-0 and vti11 seedlings grown on solvent control (DMSO) (n=23 for Col-0, n=17 for vti11) or 20 µM VAC1 (n=17 for Col-0, n=15 for vti11) plates. Statistical significance was determined by two-way ANOVA with Sidaks’s multiple comparisons test (Col-0: ****p<0.0001, vti11: ***p=0.0006). (H) Boxplots show relative growth of Col-0 and vti11 on 20 µM VAC1. Statistics is based on two-way ANOVA and multiple comparison analysis (Col-0: ****p<0.0001, vti11: ***p=0.0001). Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values. Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values.

We, accordingly, propose that VAC1 primarily interferes with vesicle fusion at the tonoplast downstream of heterotypic vesicle tethering complex. In agreement, VAC1-treated wild-type cells remarkably resembled the vacuolar morphology of untreated vti11 mutants (Figure 4E) and, hence, we subsequently addressed the contribution of vacuolar SNARE VTI11 to the VAC1 effect. We concluded that VAC1 attenuates root growth rates in a VTI11-dependent manner, because when compared to wild-type seedlings the vti11 mutants were less sensitive to VAC1 (Figure 4E–H). On the other hand, VAC1 application still induced alterations in VAMP711 localization (Figure 4E) and also root growth remained at least partially inhibited (Figure 4F–H) in the vti11 mutant background. We accordingly conclude that VAC1 interferes with the redundant, SNARE-dependent membrane delivery to the vacuole. Accordingly, we propose that VTI11 is not the sole or main target of VAC1, assuming functional redundancy at the level of vacuolar SNAREs and/or other accessory molecular targets.

Considering that VAC1 interferes with vacuolar SNARE-dependent vesicle trafficking to the vacuole, this pharmacological tool subsequently allowed us to address the importance of vesicle trafficking toward the vacuole in elongating cells. We next utilized VAC1 to visualize the rate of endocytic trafficking to the vacuole in early and late meristematic as well as early and late elongation zones. We therefore used plasma membrane marker BRI1-GFP and FM4-64-based pulse labelling of plasma membranes (Figure 5—figure supplement 1A-D; Figure 5A–I). Elongating cells displayed higher BRI1-GFP and FM4-64 uptake into VAC1 bodies when compared to meristematic cells (Figure 5—figure supplement 1C, D and Figure 5A and B). This finding implies either meristematic and elongating cells to differ in endocytic membrane trafficking or alternatively inhomogeneous VAC1 activity/uptake. To further address the possibility of altered endomembrane sorting during cellular elongation, we similarly used the well characterized inhibitor BFA to block membrane recycling and sorting toward the vacuole at the level of the TGN (Kleine-Vehn et al., 2008). In line with the results from VAC1 treatments, BFA bodies also showed greater FM4-64 signal accumulation in cells of the elongation zone when compared to cells of the meristem (Figure 5C and D). We, hence, propose that BFA and VAC1 visualize enhanced membrane sorting from the plasma membrane toward the vacuole during cellular elongation.

Figure 5 with 1 supplement see all
Endocytic trafficking is enhanced at the onset of cellular elongation.

(A) Representative maximum z-projection (scale bar: 50 µm) of 6-day-old Col-0 roots. Seedlings were pre-treated with 10 µM VAC1 for 30 min before staining in 4 µM FM4-64 (red) for 30 min. Seedlings were then de-stained for 1.5 hr and subsequently treated with 10 µM VAC1 for 3 hr before image acquisition. The treatments and the staining were conducted in liquid medium. (B) Boxplots showing mean gray values of FM4-64 fluorescence signal in VAC1 bodies in early (n=12) and late (n=12) meristematic cells as well as in the early (n=11) and late (n=8) elongation zone. One-way ANOVA with Tukey’s multiple comparisons test (b: p=0.0002, c: p<0.0001). (C) Representative maximum z-projection (scale bar: 50 µm) of 6-day-old Col-0 roots. Seedlings were pulse-stained in 4 µM FM4-64 (red) for 5 min and subsequently transferred to 25 µM BFA for 30 min prior to image acquisition. Staining and treatment was done in liquid medium. (D) Boxplots showing mean gray values of FM4-64 fluorescence signal in BFA bodies in early (n=9) and late (n=9) meristematic cells as well as in the early (n=9) and late (n=9) elongation zone. One-way ANOVA with Tukey’s multiple comparisons test (b to a [early meristem]: p=0.0123, c to a: p<0.0001, c to b: p=0.0178). (E) 3D reconstructions of plasma membrane (pUBQ10::NPSN12-YFP, green) and BFA bodies (red) in early (left) and late (right) meristematic cells. Six-day-old pUBQ10::NPSN12-YFP seedlings were pulse-stained with 4 µM FM4-64 for 5 min and subsequently transferred to 50 µM BFA for 2 hr prior image acquisition. Staining and treatment was done in liquid medium. Scale bars: 5 µm. (F) Boxplots depict relative increase of cell and BFA body volume (n=10, respectively) and relative increase of cell and BFA body surface area (n=10, respectively) in late meristematic cells when compared to early meristematic cells. Student’s t-test (*p=0.0217). (G) Representative images of early and late meristematic cells and cells of the early and late elongation zone. Six-day-old pUBQ10::YFP-VAMP711 seedlings were pulse-stained with 4 µM FM4-64 for 5 min and subsequently de-stained in liquid medium for 3 hr before image acquisition. Overlay of YFP-VAMP711 (magenta, upper panel) and FM4-64 (gray, middle panel) fluorescence signals is shown in the lower panel. Scale bar: 10 µm. Note: Images are assembled on white background. (H) Representative scatter plots depict co-localization of YFP-VAMP711 and FM4-64 in cells of the early and late meristem and in the early and late elongation zone. (I) Boxplots show Rtotal of co-localization between YFP-VAMP711 and FM4-64 in early (n=9) and late (n=9) meristematic cells and in the early (n=9) and late (n=9) elongation zone. One-way ANOVA with Tukey’s multiple comparisons test (b: p<0.0001). Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values. Boxplots: Box limits represent 25th percentile and 75th percentile; horizontal line represents median. Whiskers display min. to max. values. Data points are individual measured values.

Due to the steady surface increase, elongating cells may intrinsically show enhanced vesicle trafficking to the vacuole, even if endocytic membrane delivery would remain constant. Hence, we next performed whole-cell 3D reconstructions of early as well as late meristematic cells and observed that these cells displayed roughly a doubling of surface area and volume (Figure 5E and F). Compared to the cellular size increase, the volume of FM4-64 containing BFA bodies showed a much stronger relative size increase (Figure 5E and F). Considering that the volume of the BFA compartment is filled with small vesicles (Geldner et al., 2001), we anticipate that cellular surface enlargement alone does not fully account for the observed alterations in intracellular dynamics. We, hence, rather conclude that endocytic membrane sorting toward the vacuole undergoes reprogramming in elongating cells.

Next, we tested whether membrane delivery from the plasma membrane to the vacuole is enhanced during cellular elongation when we do not apply vesicle trafficking inhibitors. To this end, we quantified the co-localization of FM4-64 with the tonoplast marker YFP-VAMP711 in meristematic and elongating cells (Figure 5G). Three hours after the relatively equal pulse staining of FM4-64 (Figure 5—figure supplement 1A, B), the co-localization between FM4-64 and YFP-VAMP711 at the vacuolar membrane was significantly lower in the meristematic zone when compared to the elongation zone (Figure 5H, I). Taken together, these results indicate that endocytic membrane sorting from the plasma membrane toward the vacuole undergoes significant rate changes during cellular elongation.

Based on our data, it is conceivable that endocytic trafficking impacts cell expansion via its contribution to vacuolar size. To approach this, we initially quantified cell and vacuole surface areas of (early and late) meristematic and (early and late) elongating cells using 3D reconstructions. We found that the area increase of both cell and vacuole surface is remarkably similar during the course of cell expansion (Figure 6A left panel, B). To address this coordinative mechanism, we applied VAC1 and employed whole-cell 3D reconstructions (Figure 6A right panel). Within only 2.5 hr, VAC1 induced an already detectable imbalance of cell and tonoplast surface size (Figure 6C). This result indicates that intensified membrane delivery is not only required at the plasma membrane, but similarly occurs at the tonoplast during cellular expansion. Accordingly, not only actin- and myosin-dependent vacuolar unfolding (Scheuring et al., 2016), but also membrane delivery greatly contributes to vacuolar size control.

Coordinated surface increase at plasma membrane and tonoplast during cell elongation.

(A) 3D reconstructions of propidium iodide (PI)-stained cell walls (gray) and BCECF-stained vacuoles (magenta) in the early and late meristem and in the early and late elongation zone. Scale bars: 5 µm. (B) and (C) Boxplots show cell and vacuole surface areas in the defined zones and indicated treatments. Six-day-old Col-0 seedlings were treated with solvent control (dimethyl sulfoxide [DMSO], n=8) or 10 µM VAC1 (n=9) for 2.5 hr in liquid medium. Student’s t-test (early meristem: **p=0.0018, late meristem: **p=0.0073, early elongation: *p=0.0423). (D) Representative images (scale bars: 100 µm) of maximum z-projections of 6-day-old pUBQ10::NPSN12-YFP (yellow) seedlings. Seedlings were mounted on agar blocks containing solvent control (DMSO) or 20 µM VAC1 in chamber slides. t0 represents the first and t165 the last image acquisition timepoint, given in minutes after mounting. Note: Images are assembled on black background. (E) Boxplots show root tip displacement of DMSO (n=18) and VAC1 (n=19) treated samples at the end of the 165 min time frame. Student’s t-test (****p<0.0001). (F) Representative images (scale bars: 25 µm) of maximum z-projections of 6-day-old pUBQ10::NPSN12-YFP (yellow) seedlings. Seedlings were mounted on agar blocks containing solvent control (DMSO) or 20 µM VAC1 in chamber slides. t0 represents the first and t165 the last image acquisition timepoint, given in minutes after mounting. Note: Images are assembled on black background. (G) Boxplots show cell lengths of DMSO (n=15) and VAC1 (n=15) treated samples at the beginning and the end of the 165 min time frame. One-way ANOVA with Tukey’s multiple comparisons test (b to a: p<0.0001, b to c: p=0.0013, c to a: p≤0.0012).

Our data propose that the dynamic regulation of membrane sorting from the plasma membrane toward the vacuole is substantially contributing to the vacuole size increase, which could have an impact on rapid cell expansion rates. To address this, we used VAC1 to temporally block the vesicle fusion to the tonoplast during cell expansion. Accordingly, we mounted Col-0 seedlings in VAC1 or solvent (control) containing slide chambers and immediately inspected root growth under these conditions using confocal microscopy. VAC1 induced without a further delay a slower root growth (tip displacement) (Figure 6D and E), which correlated with reduced cellular expansion rates (Figure 6F and G) when compared to the solvent control. Accordingly, we conclude that vesicle trafficking to the vacuole is required for fast cellular expansion rates. In summary, our set of data suggests that endocytic trafficking toward the vacuole coordinates cellular and vacuolar surface increase, enabling rapid cell enlargements in plants.

Discussion

Here, we isolated and characterized VAC1, which interferes with the final steps of membrane delivery to the tonoplast. The molecular target of VAC1 remains currently unknown, but the structure-activity assessment suggests that the entirety of the compound is contributing to its activity, possibly insulating a complex mode of VAC1 action. Our data suggest that VAC1 disrupts SNARE-dependent, heterotypic vesicle delivery to the tonoplast. It is tempting to speculate that VAC1 interferes with possibly several subunits of the vacuolar SNARE complex or yet to be identified accessory proteins. Accordingly, in-depth characterization of the mode of VAC1 action could reveal intermolecular mechanisms of vesicle fusion at the tonoplast.

We used VAC1 to visualize the rate of vesicle trafficking toward the vacuole. Mechanistically, our data suggests that endocytic trafficking is accelerated at the onset of elongation, suggesting that plasma membrane-derived vesicles at least in part fuel the size increase of the vacuole. We propose that cell and vacuolar surface enlargements are thereby coordinated and pinpoint the importance of vacuolar size control for cell expansion and overall root growth. It remains to be seen how plant cells molecularly reprogram and monitor the rate of endocytic trafficking for ensuring fast cellular elongation. Not only VAC1 but also BFA treatments visualized higher rates of endocytic membrane accumulation in expanding cells. We, therefore, assume that the proposed reprogramming of membrane sorting occurs upstream or at the level of the TGN. The accelerated endocytic membrane flow toward the vacuole could relate to rate changes in endocytosis, but alternatively also the endosomal sorting machinery could shift the balance between endocytic recycling (back to the plasma membrane) and membrane drainage toward the vacuole.

The enhanced delivery of plasma membrane-derived membranes toward the vacuole during cell expansion seems somewhat counterintuitive, because it appears to slow down the theoretically possible cell surface increase. In addition, a high degree of membrane recycling at the plasma membrane is needed to fuel the high demand for cell wall synthesis (Robert et al., 2005; Ketelaar et al., 2008; Ebine and Ueda, 2015) and, hence, draining membranes toward the vacuole may even constrain the crucial cell wall remodeling during cell expansion. On the other hand, this vesicle trafficking-based compromise model for enlargements of plant cells allows the steady increase in vacuolar occupancy in dynamically expanding cells, which enables growth with little de novo synthesis of cytosolic components. This coordinated membrane flow mechanism is able to synchronize cell and vacuolar surface increase, which is likely the reason why plant cells excel in the speed of cellular expansion.

Materials and methods

Plant material and growth conditions

Request a detailed protocol

Experiments were carried out in Arabidopsis thaliana (Col-0 ecotype). The following plant lines were described previously: vti11 (zigzag) (Yano et al., 2003), pUBQ10::VAMP711-YFP (WAVE9Y) (Geldner et al., 2009), pUBQ10::pHGFP-VTI11 (Takemoto et al., 2018), pSYP22::SYP22-GFP in syp22 background (Uemura et al., 2010), pBRI1::BRI1-GFP (Kleine-Vehn et al., 2008), pPGP19::PGP19-GFP (Dhonukshe et al., 2008), pUBQ10::NPSN12-YFP (WAVE131Y) (Geldner et al., 2009), CLC-GFP (Ito et al., 2011), p35S::Lifeact-venus (Era et al., 2009), pUBQ10::GOT1-YFP (WAVE18Y) (Geldner et al., 2009), pVHAa1::VHAa1-GFP (Dettmer et al., 2006), pVHAa3::VHAa3-GFP (Dettmer et al., 2006), pUBQ10::RABF2a-YFP (WAVE7Y) (Geldner et al., 2009), p35S::SYP21-GFP (Robert et al., 2008), amiR vps3 (Takemoto et al., 2018), amiR vps39 (Takemoto et al., 2018), pUBQ10::YFP-VAMP711 amiR vps3 (Takemoto et al., 2018). vti11 pUBQ10::YFP-VAMP711 was obtained by crossing. Seeds were stratified at 4°C for 2 days in the dark and were grown on vertically orientated ½ strength Murashige and Skoog (MS) medium plates containing 1% sucrose under a long-day regime (16 hr light/8 hr dark) at 20–22°C.

Chemicals

All chemicals were dissolved in dimethyl sulfoxide (DMSO) (Duchefa, NL, cat. no. D1370) and applied in solid or liquid ½ MS medium. FM4-64 (cat. no. T3166) was obtained from Invitrogen/Thermo Fisher Scientific (Waltham, MA), PI (cat. no. 81845) and DEX (cat. no. D1756) from Sigma (St Louis, MO), 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) (cat. no. 51012) from Biotium (Fremont, CA), Wortmannin (WM) (cat. no. HY-10197) from MedChemExpress (Princeton, NJ) and BFA (cat. no. J62340) from Alfa Aesar (Tewksbury, MA). VAC1 (N’-[(2,4-dibromo-5-hydroxyphenyl)methylidene]-2-phenylacetohydrazide, ID 5326213), VAC1A1 (2-phenyl-N'-(2,4,6-tribromo-3-hydroxybenzylidene)acetohydrazide, ID 5326215), VAC1A2 (N'-(2-bromo-5-hydroxybenzylidene)-2-phenylacetohydrazide, ID 5326212), VAC1A3 (N'-(2,4-dibromo-5-hydroxybenzylidene)-2-(1-naphthyl)acetohydrazide, ID 5575792), VAC1A4 (N'-(2,4-dibromo-5-hydroxybenzylidene)-2-hydroxy-2-phenylacetohydrazide, ID 5568198), and VAC1A5 (N'-(3,5-dibromo-4-hydroxybenzylidene)-2-phenylacetohydrazide, ID 5326129) were obtained from ChemBridge/Hit2Lead (San Diego, CA) (https://www.hit2lead.com/search.asp?db=SC).

VAC1

Request a detailed protocol

VAC1 used in this study was either obtained from ChemBridge/Hit2Lead (San Diego, CA) (https://www.hit2lead.com/search.asp?db=SC, ID 5326213) or synthesized in-house as described in VAC1 synthesis (see below). We performed LC-MS analysis to determine the similarity of synthesized and commercially available VAC1. HPLC chromatograms and MS spectra of the synthesized VAC1 were identical to the ones of the reference VAC1. The purity of the synthesized VAC1 (99.2%) was higher than that of the purchased VAC1 (98.2 %). The IUPAC (International Union of Pure and Applied Chemistry) nomenclature was adapted to the VAC1 molecule (N’-[(2,4-dibromo-5-hydroxyphenyl)methylidene]-2-phenylacetohydrazide) throughout the manuscript, however an alternative name for VAC1 (N'-(2,4-dibromo-5-hydroxybenzylidene)-2-phenylacetohydrazide) is used by the manufacturer Chembridge/Hit2Lead.

VAC1 synthesis

Preparative route to VAC1.

Reaction conditions: (a) EtOH, cat. Fe2(SO4)3/H2SO4, reflux, 3.5 hr, 87%; (b) N2H4·nH2O (50–60% N2H4), MeOH, rt, 24–72 hr, 68%; (c) Br2 (2.12 eq.), DCM, rt, 22 hr, 64%; (d) EtOH, reflux, 3 hr, 76%.

Experimental

Request a detailed protocol

Starting reagents 3-hydroxybenzaldehyde 4 and hydrazine hydrate were purchased from Sigma-Aldrich (cat. no. H19808 and 225819, respectively) and PAA 1 was purchased from Merck (cat. no. P16621). These reagents were used as received without further purification.

Synthesis of 5 was accomplished by following the respective procedure in Kallman et al., 2014, and 2 was prepared adapting Liang et al., 2004.

TLC was run on Silica 60 glass plates (Merck). Low-resolution mass spectra were obtained on an Agilent 6340 Ion Trap instrument. For m/z values, the most abundant isotope of the isotope distribution is reported.

Ethyl phenylacetate (2)

Request a detailed protocol

A solution of 1 (4.08 g, 30 mmol) in abs. ethanol (150 ml) containing 50 mg of equimolar mixture of iron (III) sulfate and concentrated sulfuric acid was held at reflux for 3.5 hr until no starting material was visible on TLC. Ethanol was distilled off and the residue was dissolved in DCM, washed with saturated aqueous solution of sodium bicarbonate, dried with anh. magnesium sulfate, stripped of solvent and purified by distillation in vacuo (bp 105–108°C at 13 Torr) to afford 2 as colorless liquid. Yield: 4.3 g (87%).

PAA hydrazide (3)

Request a detailed protocol

Hydrazine hydrate (2.81 ml, 45–54 mmol) was added dropwise to a solution of 2 (4.0 g, 24 mmol) in methanol (28 ml) and the mixture was agitated at rt for 24–72 hr until complete consumption of the ester (TLC). Excess solvent was removed by distillation and, upon cooling, the residue crystallized as long thin needles. The mass was crushed, washed on a glass filter with cold water, and recrystallized from 90% aq. ethanol. Yield: 3.0 g (82%) before and 2.5 g (68%) after recrystallization.

2,4-Dibromo-5-hydroxybenzaldehyde (5)

Request a detailed protocol

To a solution of 4 (1.0 g, 8.2 mmol) in DCM (10 ml) was slowly added elemental bromine (2.78 g 17.4 mmol) and the reaction mixture was stirred at rt for 22 hr. The reaction progress was monitored by HPLC. Upon completion, the reaction was quenched by dropwise addition of 15% aq. sodium thiosulfate (4.8 ml), stirred at rt for 1 hr and filtered. The filter cake was washed with water (2 × 5 ml) and dissolved in acetic acid (6.4 ml) at 95°C. After cooling to 50°C, water (3.4 ml) was added dropwise to a stirred solution, resulting in crystallization of the product. The crystals were stirred for 4 hr at 15°C, filtered, washed with water (2 × 5 ml) and dried. Yield: 1.68 g (73%). The product (pink crystals) was additionally purified by vacuum sublimation, affording 5 as white crystalline powder (1.48 g, 64%). m/z (ESI-): 278.8 [M-H]-.

N’-[(2,4-dibromo-5-hydroxyphenyl)methylidene]-2-phenylacetohydrazide (VAC1)

Request a detailed protocol

A mixture of 3 (0.53 g, 3.53 mmol) and 1.02 eq. of 5 in 9.6 ml of ethanol was held at reflux for 3 hr until complete consumption of the starting material (TLC). Upon cooling, a copious white precipitate of the title compound was formed. It was filtered, washed with petroleum ether, and dried. The final product was additionally purified by recrystallization from ethanol. Yield 1.1 g (76%). m/z (ESI-): 410.9 [M-H]-.

UHPLC-UV-(SIR)MS

Request a detailed protocol

To determine the conversion of VAC1 into free PAA, 6-day-old seedlings were treated with DMSO or 10 µM VAC1 for 2.5 hr in liquid medium and then flash-frozen in liquid nitrogen. The extraction and quantification methods for both analytes were optimized based on the previously described method (Pařízková et al., 2021). Briefly, the plant samples (~30 mg FW) were extracted into 100% methanol and purified using liquid-liquid extraction method into 900 µl of extraction solution (methanol:H2O: hexane – 1:1:1), evaporated and dissolved in 50 µl of 100% methanol. Two µl of each sample were injected onto the reversed-phase column (Kinetex C18 100A, length 50 mm, diameter 2.1 mm, particle size 1.7 μm; Phenomenex, Torrance, CA) and analyzed by an Acquity UPLC H-Class System (Waters, Milford, MA) coupled with an Acquity PDA detector (scanning range 190–400 nm with 1.2 nm resolution) and a single quadrupole mass spectrometer QDa MS (Waters MS Technologies, Manchester, UK) equipped with an electrospray interface (ESI). The analytes were eluted from the column within 9-min-linear gradient of 10:90 to 95:5 A:B using 0.1% acetic acid in methanol (A) and 0.1% acetic acid in water (B) as mobile phases at a flow rate of 0.5 ml·min-1 and column temperature of 40°C. After every analysis, the column was washed with 95% methanol and then equilibrated to initial conditions (1.0 min). Both analytes were detected by selected ion recording (SIR) using the positive and negative electrospray modes (ESI+ and ESI-) as follows: VAC1 detected as [M+H]+, m/z 411.0, and PAA as [M-H]-, m/z 135.0. The MS settings were optimized as follows: source temperature, 120°C; desolvation temperature, 600°C; capillary voltage, 0.8 kV. Chromatograms were processed by MassLynx V4.2 software (Waters) and quantification was performed from external calibration using a recovery factor. Additionally, M1-M3 metabolites of VAC1 were detected using a full-scan mode (m/z 50–1000) operated in ESI- with post-data acquisition extraction of ion chromatograms for m/z 425.0, 571.0, and 613.0 for M1, M2, and M3, respectively. To determine the distribution of VAC1 metabolites, the UV chromatograms were extracted for λmax (291 nm) and the percentage representation of metabolites in plant extracts was determined by the integration of peak areas in respective retention times (M1, 4.37 min; M2, 5.26 min; M3, 5.39 min; VAC1, 6.65 min).

UHPLC-HRMS

Request a detailed protocol

Samples of DMSO- and VAC1-treated plants were prepared as described above (UHPLC-SIR-MS). Plant extracts (2 µl) were injected onto a Kinetex C18 column and separated using chromatographic conditions as described above. The HRMS analysis was achieved by a hybrid Q-TOF tandem mass spectrometer Synapt G2-Si (Waters MS Technologies) as described previously (Buček et al., 2018). Briefly, the effluents were introduced into the HRMS instrument (ESI-; capillary voltage, 0.75 kV; source offset, 30 V; desolvation/source temperature, 550/120°C; desolvation/cone gas flow, 1000/50 l·hr-1; LM/HM resolution, 2.8/14.75; ion energy 1/2, 0.5/1.0 V; entrance/exit voltages, 0.5 V; collision energy, 6 eV). The determination of exact mass was performed by the external calibration using lock spray technology and a mixture of leucine/encephalin (1 ng·μl-1) in an acetonitrile and water (1:1) solution with 0.1% formic acid as a reference. Data acquisition was performed in full-scan mode (50–1000 Da) with a scan time of 0.5 s and all analytes were detected as [M-H]- in the MS spectrum. All data were processed using MassLynx 4.1 software (Waters). The accurate masses of VAC1 metabolites were calculated and then used to determine the elemental composition of the analytes with a fidelity ranging from 1.6 to 2.6 ppm. The VAC1 metabolites were identified based on correlation of the theoretical monoisotopic weights of deprotonated forms [M-H]- and detected accurate masses of each precursor of M1-M3 as well as based on the presence of two bromine stable isotopes (79Br and 81Br) in HRMS spectrum resulting in the characteristic isotope pattern of dibromo derivatives.

Phenotype analysis

Request a detailed protocol

Vacuolar occupancy and vacuole surface area as well as cell surface area and BFA body volume were quantified in 6-day-old seedlings. For 3D reconstructions of cells, confocal images were processed using Imaris (vacuolar occupancy of cells, BFA bodies, vacuole, and cell surface) as described previously (Dünser et al., 2019). BCECF staining was performed as described previously (Scheuring et al., 2015). For the analysis of main root growth, 6-day-old seedlings were used, unless indicated otherwise. For the analysis of main root growth of amiR vps3 and amiR vps39 seedlings were grown for 4 days before transfer to DMSO, DEX, VAC1, or DEX + VAC1 and then grown for additional 3 days. Plates were scanned and root length assessed using ImageJ. Selection of pUBQ10::YFP-VAMP711 amiR vps3 seedlings for confocal microscopy analysis was done as described previously (Takemoto et al., 2018).

Confocal microscopy

Request a detailed protocol

For image acquisition, a Leica TCS SP5 (DM6000 CS) or a Leica SP8 (DMi8) confocal laser scanning microscope, equipped with a Leica HCX PL APO CS 63×1.20 water-immersion objective, was used. GFP and BCECF were excited at 488 nm (fluorescence emission: 500–550 nm), YFP and FM4-64 at 514 nm (fluorescence emission YFP: 525–578 nm; fluorescence emission FM4-64: 670–790 nm), and PI at 561 nm (fluorescence emission: 644–753 nm). Roots were mounted in PI solution (0.02 mg/ml) for the counterstaining of cell walls. Z-stacks were recorded with a step size of 420 nm (for 3D reconstructions) or 1.5 µm (for maximum projections of VAC1 bodies).

3D reconstruction of cells

Request a detailed protocol

Imaris 8.4.0 and 9.0 were used for the reconstruction of cell, vacuole, and BFA body volumes. Based on the PI channel, every third slice of the z-stack was utilized to define the cell borders using the isoline, magic wand, or manual (distance) drawing functions in the manual surface creation tool. After creating the surface corresponding to the entire cell, a masked channel (based on BCECF or FM4-64) was generated by setting the voxels outside the surface to 0. Subsequently, a second surface (based on the masked BCECF or FM4-64 channel) was generated automatically with the smooth option checked. Surface detail was set to identical values within each region (early and late meristem and early and late elongation zone) to ensure comparability of obtained surface area values. The obtained surface was visually compared to the underlying BCECF or FM4-64 channel and, if necessary, the surface was fitted to the underlying signal by adjusting the absolute intensity threshold slider. Finally, volumes and surface areas of both surfaces were extracted from the statistics window.

Sample preparation for electron microscopy

Request a detailed protocol

Around 3 mm of the Arabidopsis root tips were sectioned and immediately submerged in fixative solution containing 2.5% glutaraldehyde and 4% p-formaldehyde in MTSB buffer. Samples were vacuum infiltrated for 15 min at room temperature, fixed for 4 hr at room temperature, and kept in fixative overnight at 4°C. After five times washing (10 min each) in MTSB buffer, samples were post-fixed in 1% osmium tetroxide in H2O on ice for 4 hr with shaking. Samples were washed with water four times (5 min each), and in bloc stained for 2 hr with 2% uranyl acetate in water. After washing three times with water (5 min each), samples were dehydrated in EtOH graded series (30%–50%−70%–80%−90%–95%, 15 min each). Fully dehydration was achieved by incubating the samples twice in 100% EtOH for 30 min followed by incubation in acetone twice (30 min each). Samples were gradually embedded in resin by incubating them in acetone: Agar 100 mixtures (3:1, 1:1, 1:3–8 hr each) and finally in pure Agar 100 resin three times (8 hr each). The root tips were flat embedded according to Kolotuev, 2014, and the resin was polymerized overnight at 60°C. The flat embedded samples were mounted on resin blocks and 70 nm longitudinal sections of root-tips were obtained with a Reichert-Jung Ultracut-E ultramicrotome and collected on slot grids. Sections were contrasted with 2% uranyl acetate and lead citrate according to Reynolds, 1963, and examined in a Hitachi 7800 TEM (100 kV) equipped with an EMSIS Xarosa camera.

Small molecule screen

Request a detailed protocol

Roughly 20 pUBQ10::YFP-VAMP711 seeds were germinated in 12-well plates with each well containing 1.5 ml liquid ½ MS+ supplemented with solvent control (DMSO) or 50 µM of small molecules. Four-day-old seedlings were screened for enhanced fluorescence signal using a fluorescence binocular. Twelve hit compounds were further examined using confocal microscopy (see Figure 2—figure supplement 1 and Figure 2—figure supplement 2). The small molecules used for the primary screen were identified as inhibitors of tobacco pollen germination (Drakakaki et al., 2011).

Quantification of VAC1 and BFA bodies

Request a detailed protocol

Maximum z-projections of three to eight slices (step size 1.5 µm) and an ROI of 5 µm × 5 µm were used for quantification of the FM4-64 fluorescence signal. Four to ten roots were quantified with three bodies per root.

Quantification of intracellular BRI1-GFP fluorescence signal

Request a detailed protocol

Maximum z-projections of three to six slices (step size 1.5 µm) and an ROI of 10 µm × 5 µm were used for quantification of the intracellular BRI1-GFP fluorescence signal. Eight roots were quantified, using two early meristematic and two late meristematic cells per root.

Co-localization analysis

Request a detailed protocol

Quantification of co-localization was done using ImageJ. Two to three atrichoblast cells in the specified regions were used per root. The area was marked and cropped (Image > Crop), subsequently the surrounding of the cell was cleared (Edit > Clear outside) and the channels were split (Image > Color > Split channels). Co-localization (Analyze > Co-localization > Co-localization threshold) was analyzed and the values for Rtotal as well as the scatter plots were extracted.

Time course experiments

Request a detailed protocol

Seedlings were mounted on agar blocks containing solvent control DMSO or VAC1. The agar blocks were placed in chamber slides and subsequently installed on an inverted confocal microscope (Leica SP8). Z-stacks (three to five slices, step size: 1.5 µm) were acquired every 10 min and maximum z-projections used for quantification of root tip displacement and cell length.

Cell length analysis

Request a detailed protocol

Six-day-old seedlings were used for cell length measurements of differentiated atrichoblasts (Figure 1C and D). Fully differentiated cells were identified as described previously (Löfke et al., 2015).

Statistical analysis and reproducibility

Request a detailed protocol

All graphs were generated using GraphPad Prism 9. Each experiment was repeated at least twice independently with similar results. The depicted data are the results from one representative experiment.

Data availability

Source Data file contains the numerical data used to generate the figures.

References

  1. Book
    1. Scheuring D
    2. Schöller M
    3. Kleine-Vehn J
    4. Löfke C
    (2015) Vacuolar staining methods in plant cells
    In: Estevez JM, editors. In Plant Cell Expansion. New York, NY: Humana Press. pp. 83–92.
    https://doi.org/10.1007/978-1-4939-1902-4_8

Decision letter

  1. Heather E McFarlane
    Reviewing Editor; University of Toronto, Canada
  2. Detlef Weigel
    Senior Editor; Max Planck Institute for Biology Tübingen, Germany
  3. Takashi Ueda
    Reviewer; National Institute for Basic Biology, Japan

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Endocytic Trafficking Promotes Vacuolar Enlargements for Fast Cell Expansion Rates in Plants" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Detlef Weigel as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Takashi Ueda (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

The reviewers agree that the paper will be interesting to a wide audience, but have highlighted three key points that are essential to address with revisions:

1. Conduct live cell imaging of endocytic markers (e.g. CME components) and/or an additional cargo to support the claim that endocytic rate is different between meristematic and elongating cells, according to suggestions made by Reviewers 1 and 2.

2. Revise the discussion around Figures 5 and 6 to clarify why cell surface area is being compared to BFA body volume (as Reviewer 2 suggests) and/or provide a more rigorous analysis (as Reviewers 1 and 3 suggest).

3. Revise the results and discussion surrounding the WM experiment as indicated by Reviewers 2 and 3. An additional PI3K inhibitor (e.g. LY294002) and/or treatment with a PI4K inhibitor (e.g. PAO) as a negative control would further support these claims.

The reviewers have also raised a number of additional points that are worth consideration, and their reviews are included below.

Reviewer #1 (Recommendations for the authors):

1. From the perspective of cell wall synthesis, it's unsurprising that increased endocytic trafficking is correlated with cell expansion since the volume of the cell wall must increase much greater than the surface area of the plasma membrane during elongative growth. I think the discussion could be improved by incorporating this idea, e.g. the work of Anne-Mie Emons (Ketelaar et al. 2008).

2. FM4-64 quantification is difficult because it is challenging to ensure that uptake is equal across cells. For example, I am surprised that so many cells in the meristematic zone showed R = 0 (data points at the bottom of the boxes in Figure 5I), indicating that even after 2 hours, no FM4-64 has reached the tonoplast. This is especially puzzling because the authors have presented images of meristematic cells with clear FM4-64 labelling of the tonoplast in other figures (e.g. Figure 3A, Figure 5G). What do the authors speculate is happening in these cases?

3. Does the difference in "endocytic rate" implied by more intense FM4-64 staining and larger BFA bodies in the elongation zone, compared to the meristematic zone really imply more endocytosis? Isn't there also more FM4-64 available since there is more surface area to the cell? Then the BFA bodies might appear brighter or larger simply because there is more dye in them? This claim probably requires secondary evidence, such as quantification of density and dynamics of a clathrin marker in these different cell types, or pulse-chase of a marker (e.g. photoactivatable or photoconvertible). In particular, the authors' hypothesis makes an easily-testable prediction: that endocytic events should be more frequent at the PM in elongation-zone cells, compared to the meristematic zone, even after accounting for the change in surface area. The authors already have a CME marker, so this is a straight-forward experiment for them to conduct.

4. It is confusing to compare surface area of a large object (the cell) to volume of a small object (the BFA body) as in Figure 5F. There is a crossover point in surface area-to-volume ratios at which the ratio is >1 for a small object and <1 for a large object and I think this graph oversimplifies this important point. The volume of the BFA body is small relative to the surface area of the cell, so the FM4-64 signal in the BFA body is concentrated, compared to the signal at the cell surface. I suggest the authors present these data some other way (e.g. plot surface area vs volume for BFA bodies and cells at different stages) to better illustrate this. I also think that the results and discussion surrounding this figure should be revised to clearly articulate this issue.

5. For the wide readership of eLife, it would be great to see whether VAC1 has any activity on vacuolar shape in other plants, or whether it has any effects on the endomembrane system in mammalian or fungal cells.

6. After introducing the vti11 phenotype, the authors conclude that "compensatory mechanisms" ensure vacuolar occupancy of the cell and "safeguard cell expansion" however vti11 vti12 double mutants are lethal, indicating that genetic redundancy, not compensatory mechanisms, are at work. Please correct.

Reviewer #2 (Recommendations for the authors):

– A marker-assisted comparison between VAC1 and Concanamycin A might be useful here as the latter compound also interferes with vacuolar membrane fusion, albeit likely less specific.

– For some of the cell types analyzed in Figure 1F, the numbers are a bit low. Some additional quantifications could be performed here to strengthen the data.

– On line 89, I would rephrase the striking effect of VAC1 on vacuolar morphology as several other VACs have different, yet equally striking effects.

– The viability of the cells upon treatment with VAC1A4 would be better addressed by FDA rather than with PI.

– On line 129, as it was not tested, the authors could include the possibility that the alternative compounds do not have a similar effect because they might be less stable that the original VAC1.

– On line 134, the statement that VAC1 interferes with the localization of the SNARE complex is odd. It affects vacuolar morphology, hence obviously it affects vacuolar SNARE localization. The authors might want to rephrase this by saying that VAC1 does not specifically affect one of the components of the SNARE complex as they observe similar localizations for various subunits of this complex.

– On line 136, it is strange to use Scheuring et al., 2015 as initial reference for the styryl dye FM4-64.

– On line 144, the assessment of the trafficking of PM resident proteins to the vacuole might be better visualized using dark-treatment to visualize differential vacuolar GFP accumulation.

– On line 150, the conclusion that VAC1 does not affect endocytic trafficking or actin dynamics could be better addressed using dynamic imaging e.g. using TIRF or via rainbow-color overlays of different timepoints of confocal slices for example.

– In Figure 3F, optimally, the quantifications could be normalized instead of measured by mean grey values. The PM signal appears very different between VAC1 and VAC1 + BFA. Also, the results of VAC1 + Wm are puzzling, the PM staining of FM in panel G is very different from the one in panel E and the disappearance of BRI1 from the TGN upon the combined treatment of VAC1 and Wm is not addressed. For clarity, I would remove the Wm data as the observations cannot be accurately explained for the moment and no clear conclusion can be drawn from this.

– In supplemental Figure 5, it is unclear to me why the emphasis lies on transversal PM.

– On line 217 and following, the authors propose enhanced endocytic trafficking correlating with elongation based on the size of VAC1 compartments and BFA bodies, which are larger in elongating cells versus meristematic cells. This conclusion assumes that there are no differences between these cells in terms of compound penetrance or the differential activity of trafficking pathways other than endocytic that could affect the size of the observed bodies. This assumption could be mentioned/discussed. In order to provide stronger support for their claim that endocytic trafficking is enhanced in elongating cells versus meristematic cells, time lapse imaging of roots using PM resident markers transiently expressed via a heat shock promotor or constitutively expressed following CHX could be considered.

– On line 229-230, the authors might want to explain better their choice to measure area instead of volume for the non-expert reader.

Reviewer #3 (Recommendations for the authors):

Figure 3G. Wm also inhibits an early event of endocytosis at the PM, probably through inhibition of production of PI4P and/or PI45P2 in the plasma membrane. The result presented in Figure 3G and H should be interpreted taking this effect of WM into authors' account.

Figure 4C, G, and S4. The definition of "relative growth" is unclear. Is this a relative value of root lengths to the mean value of root length of control plants? If this is the case, variation in the control plants is not correctly reflected in the data. Authors should interpret the results based on Figure 4B and Figure 4G?

Figure 5A-D. It would be not appropriate to compare just the mean grey value to see an endocytic activity. The mean grey value could be higher in larger cells with a larger area of the PM even if the endocytic rate is constant between these cell types.

https://doi.org/10.7554/eLife.75945.sa1

Author response

Reviewer #1 (Recommendations for the authors):

1. From the perspective of cell wall synthesis, it's unsurprising that increased endocytic trafficking is correlated with cell expansion since the volume of the cell wall must increase much greater than the surface area of the plasma membrane during elongative growth. I think the discussion could be improved by incorporating this idea, e.g. the work of Anne-Mie Emons (Ketelaar et al. 2008).

We thank you for this comment. We personally reached out to the first author of the mentioned publication and discussed this aspect. Based on the published assumption, plants have a substantial need for membrane trafficking at the plasma membrane, which is likely maintained by membrane recycling between the plasma membrane and the trans golgi network/early endosome (facilitated by constitutive endocytosis and exocytosis). Accordingly, vesicle sorting to the vacuole drains membranes away from the recycling pathway, possibly affecting not only surface enlargements, but also cell wall synthesis during cellular elongation. We thank you for raising this aspect and we have modified the discussion along these lines.

2. FM4-64 quantification is difficult because it is challenging to ensure that uptake is equal across cells. For example, I am surprised that so many cells in the meristematic zone showed R = 0 (data points at the bottom of the boxes in Figure 5I), indicating that even after 2 hours, no FM4-64 has reached the tonoplast. This is especially puzzling because the authors have presented images of meristematic cells with clear FM4-64 labelling of the tonoplast in other figures (e.g. Figure 3A, Figure 5G). What do the authors speculate is happening in these cases?

We indeed propose that FM4-64 uptake is not equal along the roots, showing enhancement at the onset of elongation. The initial FM4-64 labelling of the plasma membrane was relatively constant along the root zonation (see Figure 5—figure supplement 1 A-B).

The degree of FM4-64 labeling may slightly fluctuate in individual cells (at/before onset of elongation) and experimental conditions, but was overall consistent in all our experiments. We reproducibly observe enhanced FM4-64 accumulation in BFA and VAC1 bodies in elongating cells. This effect could relate to regulations of endocytosis at the plasma membrane and/or altered membrane sorting at the trans golgi network. Both endocytosis and endosomal sorting independently contribute to the rate of endocytic membrane trafficking to the vacuole. We feel this aspect was unclear and we improved the discussion on this matter in the revised version of our manuscript.

Figure 3A display epidermal cells at the onset of elongation (note that cells are slightly longer than wide) in the transition zone of wild type seedlings, displaying FM4-64 labeling at the tonoplast. Figure 5G shows FM4-64 labeling in the transgenic pUBQ10:YFP-VAMP711 line, which may differ in some aspects compared to wild type, but does allow the assessment of colocalization. While most cells show some degree of FM4-64 at the tonoplast in late meristematic cells, we indeed captured early meristematic cells with less FM4-64 tonoplast labeling (Figure 5G). In contrast, elongating cells always showed a high degree of colocalization.

3. Does the difference in "endocytic rate" implied by more intense FM4-64 staining and larger BFA bodies in the elongation zone, compared to the meristematic zone really imply more endocytosis? Isn't there also more FM4-64 available since there is more surface area to the cell? Then the BFA bodies might appear brighter or larger simply because there is more dye in them? This claim probably requires secondary evidence, such as quantification of density and dynamics of a clathrin marker in these different cell types, or pulse-chase of a marker (e.g. photoactivatable or photoconvertible). In particular, the authors' hypothesis makes an easily-testable prediction: that endocytic events should be more frequent at the PM in elongation-zone cells, compared to the meristematic zone, even after accounting for the change in surface area. The authors already have a CME marker, so this is a straight-forward experiment for them to conduct.

We thank you for raising this point, which certainly helped us to improve the description and discussion of our work. We realized that the term endocytic trafficking towards the vacuole can be misleading. The degree of membrane flow from the plasma membrane to the vacuole indeed depends on endocytosis, but is also defined by sorting endosomes, which gate recycling of membranes back to the plasma membrane and further endocytic membrane trafficking to the vacuole. Accordingly, enhanced endocytic trafficking to the vacuole could be achieved independent of the endocytosis rates. We have improved the discussion on this matter and hope to have thereby clarified this aspect. While our work suggests a reprogramming of vesicle trafficking events at the onset of elongation, upcoming work should in depth address the underlying molecular mechanism. Considering its potential complexity, this aspect is however beyond the scope of this manuscript.

If the endocytosis rate remains constant, a cellular surface increase could indeed lead to higher uptake of endocytic dyes. The elucidation of the cellular surface is a read-out for the amount of plasma membranes. In contrast, a BFA compartment is filled with secretory vesicles (with a diameter of below 0,1 µm; see Geldner et al., 2001) and hence the BFA body surface insufficiently quantifies these membranes. We hence addressed both the relative increase in cellular surface and volume between early and late meristematic cells. While these cells roughly doubled in size, particularly the BFA volume showed a much stronger relative increase. Accordingly, we assume that the increase in cell surface does not recapitulate the increases in vacuolar membrane trafficking, suggesting that vesicle trafficking undergoes reprogramming during cellular elongation. We however certainly do not rule out that the surface increase contributes to this mechanism. We improved the depiction and discussion of this set of data.

4. It is confusing to compare surface area of a large object (the cell) to volume of a small object (the BFA body) as in Figure 5F. There is a crossover point in surface area-to-volume ratios at which the ratio is >1 for a small object and <1 for a large object and I think this graph oversimplifies this important point. The volume of the BFA body is small relative to the surface area of the cell, so the FM4-64 signal in the BFA body is concentrated, compared to the signal at the cell surface. I suggest the authors present these data some other way (e.g. plot surface area vs volume for BFA bodies and cells at different stages) to better illustrate this. I also think that the results and discussion surrounding this figure should be revised to clearly articulate this issue.

In the revised version, we show now the relative surface and volume increase for the cell and for the BFA compartment. In both cases, the relative increase of the BFA compartment (surface and volume) exceeds the relative size increase of the cell. Considering the accumulation of the secretory vesicles in the BFA compartment, we propose that the relative increase in BFA compartment size is not a mere consequence of cell size increase. We improved the discussion surrounding this finding in the revised version of the manuscript.

5. For the wide readership of eLife, it would be great to see whether VAC1 has any activity on vacuolar shape in other plants, or whether it has any effects on the endomembrane system in mammalian or fungal cells.

We are indeed interested in testing VAC1 in other plants, mammalian and fungal cells and also have ongoing collaborations with other labs on this topic. However, this aspect needs to be thoroughly addressed and is beyond the scope of this manuscript.

6. After introducing the vti11 phenotype, the authors conclude that "compensatory mechanisms" ensure vacuolar occupancy of the cell and "safeguard cell expansion" however vti11 vti12 double mutants are lethal, indicating that genetic redundancy, not compensatory mechanisms, are at work. Please correct.

We agree that VAC1 likely overcomes genetic redundancy. On the other hand, we do not know whether the roundish vacuoles in vti11 mutants are due to redundant VTI12 function or whether this relates to other compensatory mechanisms. We, however, followed your suggestion and simplified our statement.

Reviewer #2 (Recommendations for the authors):

– A marker-assisted comparison between VAC1 and Concanamycin A might be useful here as the latter compound also interferes with vacuolar membrane fusion, albeit likely less specific.

A similar analysis has been performed (see Dettmer et al., 2006) and FM4-64 + ConcA looks quite different from FM4-64 + VAC1. In contrast to VAC1, ConcA also impacts on the integrity of the trans golgi network.

– For some of the cell types analyzed in Figure 1F, the numbers are a bit low. Some additional quantifications could be performed here to strengthen the data.

Detailed z-stack imaging and 3D reconstructions of different cell types are unfortunately very elaborate and time consuming. It is therefore hardly feasible to reach a very high sample size during a single imaging session. The data was highly reproducible in our independent experiments. Instead of pooling distinct biological replicates, we preferred to show a representative experiment with a sample size between 7 and 16 reconstructed cells for each category.

– On line 89, I would rephrase the striking effect of VAC1 on vacuolar morphology as several other VACs have different, yet equally striking effects.

We agree with your suggestion and rephrased our wording.

– The viability of the cells upon treatment with VAC1A4 would be better addressed by FDA rather than with PI.

Thank you for the suggestion. When characterizing VAC1A4 in more detail, we will certainly consider using fluorescein diacetate as a viability staining in our future experimental set up.

– On line 129, as it was not tested, the authors could include the possibility that the alternative compounds do not have a similar effect because they might be less stable that the original VAC1.

We fully agree. In the revised version of this manuscript, we noted this possibility.

– On line 134, the statement that VAC1 interferes with the localization of the SNARE complex is odd. It affects vacuolar morphology, hence obviously it affects vacuolar SNARE localization. The authors might want to rephrase this by saying that VAC1 does not specifically affect one of the components of the SNARE complex as they observe similar localizations for various subunits of this complex.

VAC1 treatment indeed affects the morphology of the vacuole, but also induces a preferential accumulation of vacuolar SNARE proteins in aster like aggregates. We rephrased and improved the description of this aspect in the revised version of this manuscript.

– On line 136, it is strange to use Scheuring et al., 2015 as initial reference for the styryl dye FM4-64.

We included additional references in the revised version, such as:

Vida and Emr (1995): A new vital stain for visualizing vacuolar membrane dynamics and endocytosis in yeast. The Journal of Cell Biology, 128(5), 779-792.

Ueda, T., Yamaguchi, M., Uchimiya, H., and Nakano, A. (2001). Ara6, a plant-unique novel type Rab GTPase, functions in the endocytic pathway of Arabidopsis thalianaThe EMBO Journal20(17), 4730-4741.

– On line 144, the assessment of the trafficking of PM resident proteins to the vacuole might be better visualized using dark-treatment to visualize differential vacuolar GFP accumulation.

We indeed have previously established the dark assay to address the turnover of plasma membrane proteins (Kleine-Vehn et al., 2008). However, dark conditions also affect growth responses, which is hence not ideal for research on cell expansion. VAC1 treatments could become hence a useful tool to address the rate of protein trafficking towards the vacuole, which may complement dark treatments.

– On line 150, the conclusion that VAC1 does not affect endocytic trafficking or actin dynamics could be better addressed using dynamic imaging e.g. using TIRF or via rainbow-color overlays of different timepoints of confocal slices for example.

We fully agree that additional work would be needed to address the functionality of endocytosis and actin cytoskeleton in full. We improved the description of this aspect in our revised version.

– In Figure 3F, optimally, the quantifications could be normalized instead of measured by mean grey values. The PM signal appears very different between VAC1 and VAC1 + BFA. Also, the results of VAC1 + Wm are puzzling, the PM staining of FM in panel G is very different from the one in panel E and the disappearance of BRI1 from the TGN upon the combined treatment of VAC1 and Wm is not addressed. For clarity, I would remove the Wm data as the observations cannot be accurately explained for the moment and no clear conclusion can be drawn from this.

Normalization of BFA body signal to PM signal for each experiment could be troubling, because BFA may also affect the signal at the PM as it reduces recycling events. However, we assume that the appearance of discriminable BFA compartments (labelled by FM4-64 only) and the reduced signal intensity of FM4-64 in VAC1 compartments (labelled by FM4-64 and SYP21), is very suggestive that VAC1 functions downstream of the TGN. We agree that the data on Wm is not fully conclusive. We followed your suggestion and removed our Wm related data from the revised version of our manuscript.

– In supplemental Figure 5, it is unclear to me why the emphasis lies on transversal PM.

The quantification of signal intensities of transversal plasma membranes of z-stack maximum projections allows, to our opinion, more accurate/precise measurements of the respective cell type. The signal from longitudinal plasma membranes may contain information from subjacent (cortex) cells, which we hereby exclude.

– On line 217 and following, the authors propose enhanced endocytic trafficking correlating with elongation based on the size of VAC1 compartments and BFA bodies, which are larger in elongating cells versus meristematic cells. This conclusion assumes that there are no differences between these cells in terms of compound penetrance or the differential activity of trafficking pathways other than endocytic that could affect the size of the observed bodies. This assumption could be mentioned/discussed. In order to provide stronger support for their claim that endocytic trafficking is enhanced in elongating cells versus meristematic cells, time lapse imaging of roots using PM resident markers transiently expressed via a heat shock promotor or constitutively expressed following CHX could be considered.

When inspecting meristematic and elongating epidermal cells, VAC1 fully blocked in all inspected cells the FM4-64 uptake into the vacuole. Moreover, we did not detect obvious differences in VAC1 induced accumulation of vacuolar SNAREs. Accordingly, we assume that VAC1 shows similar activity in meristematic and elongating cells. Considering that BFA treatments (Figure 5C-F) as well as drug free assays (Figure 5G-I) had a similar outcome, we assume that drug penetrance is likely of lower importance.

We also followed your suggestion and addressed also the VAC1 sensitivity of plasma membrane marker pBRI1:BRI1-GFP, which also showed more intense BRI1 accumulation in elongating epidermal cells. We included this set of data and improved as suggested the overall discussion of this matter in the revised version of this manuscript.

– On line 229-230, the authors might want to explain better their choice to measure area instead of volume for the non-expert reader.

BFA compartments are comprised of accumulating vesicles and hence the surface of the compartment may not fully capture the amounts of membranes. In the revised version of this manuscript, we now provide the relative increase of the surface as well as the volume for both the cell and BFA compartment. We improved the data depiction and overall discussion on this matter. For further details, please, also see response to Reviewer 1.

Reviewer #3 (Recommendations for the authors):

Figure 3G. Wm also inhibits an early event of endocytosis at the PM, probably through inhibition of production of PI4P and/or PI45P2 in the plasma membrane. The result presented in Figure 3G and H should be interpreted taking this effect of WM into authors' account.

We agree with your concerns, which were also raised by reviewer 2. Considering that the interpretation of the data is troubling, we followed the suggestion of Reviewer 2 and removed the data in the revised version of this manuscript.

Figure 4C, G, and S4. The definition of "relative growth" is unclear. Is this a relative value of root lengths to the mean value of root length of control plants? If this is the case, variation in the control plants is not correctly reflected in the data. Authors should interpret the results based on Figure 4B and Figure 4G?

Thank you for the comment. We adapted this analysis and displayed the variation of control plants in the revised version.

Figure 5A-D. It would be not appropriate to compare just the mean grey value to see an endocytic activity. The mean grey value could be higher in larger cells with a larger area of the PM even if the endocytic rate is constant between these cell types.

We thank you for this comment. We agree that an increase in endocytic membrane flow to the vacuole may not be dependent on an increase in endocytosis rate. We improved the discussion on this matter in the revised version of the manuscript. When compared to the cell size increase, we however observed accelerated size increase of BFA compartments during cellular enlargements, rather suggesting that a reprogramming of intracellular membrane trafficking contributes to this effect. Please, also see additional comments in reply to reviewer 1 and 2.

https://doi.org/10.7554/eLife.75945.sa2

Article and author information

Author details

  1. Kai Dünser

    1. Molecular Plant Physiology (MoPP), Faculty of Biology, University of Freiburg, Freiburg, Germany
    2. Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    3. Center for Integrative Biological Signalling Studies (CIBSS), University of Freiburg, Freiburg, Germany
    Contribution
    Conceptualization, Data curation, Funding acquisition, Investigation, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8974-2072
  2. Maria Schöller

    Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    Contribution
    Data curation, Formal analysis, Investigation, Validation, Visualization
    Competing interests
    No competing interests declared
  3. Ann-Kathrin Rößling

    1. Molecular Plant Physiology (MoPP), Faculty of Biology, University of Freiburg, Freiburg, Germany
    2. Center for Integrative Biological Signalling Studies (CIBSS), University of Freiburg, Freiburg, Germany
    Contribution
    Data curation, Investigation, Validation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
  4. Christian Löfke

    Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    Contribution
    Data curation, Formal analysis, Investigation, Supervision, Validation, Visualization
    Competing interests
    No competing interests declared
  5. Nannan Xiao

    Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    Contribution
    Investigation, Validation
    Competing interests
    No competing interests declared
  6. Barbora Pařízková

    Laboratory of Growth Regulators, Institute of Experimental Botany of the Czech Academy of Sciences and Faculty of Science of Palacký University, Olomouc, Czech Republic
    Contribution
    BP conducted UHPLC-SIR-MS and UHPLC-HRMS., Data curation, Formal analysis, Investigation, Validation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
  7. Stanislav Melnik

    Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    Contribution
    Investigation, SM synthesised VAC1., Validation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
  8. Marta Rodriguez-Franco

    Cell Biology, Faculty of Biology, University of Freiburg, Freiburg, Germany
    Contribution
    Investigation, MR-F conducted transmission electron microscopy., Validation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
  9. Eva Stöger

    Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  10. Ondřej Novák

    Laboratory of Growth Regulators, Institute of Experimental Botany of the Czech Academy of Sciences and Faculty of Science of Palacký University, Olomouc, Czech Republic
    Contribution
    Data curation, Formal analysis, Funding acquisition, Investigation, ON conducted UHPLC-SIR-MS and UHPLC-HRMS., Resources, Supervision, Validation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
  11. Jürgen Kleine-Vehn

    1. Molecular Plant Physiology (MoPP), Faculty of Biology, University of Freiburg, Freiburg, Germany
    2. Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria
    3. Center for Integrative Biological Signalling Studies (CIBSS), University of Freiburg, Freiburg, Germany
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Methodology, Resources, Supervision, Validation, Writing – original draft, Writing – review and editing
    For correspondence
    juergen.kleine-vehn@biologie.uni-freiburg.de
    Competing interests
    Senior editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4354-3756

Funding

Austrian Science Fund (P 33044)

  • Jürgen Kleine-Vehn

Deutsche Forschungsgemeinschaft (470007283)

  • Jürgen Kleine-Vehn

CIBSS (EXC-2189)

  • Jürgen Kleine-Vehn

European Research Council (639478-AuxinER)

  • Jürgen Kleine-Vehn

Ministry of Education, Youth and Sports of the Czech Republic (European Regional Development Fund-Project "Plants as a tool for sustainable global development" (CZ.02.1.01/0.0/0.0/16_019/0000827)

  • Ondřej Novák

Deutsche Forschungsgemeinschaft (426849454)

  • Marta Rodriguez-Franco

Austrian Academy of Sciences (DOC fellowship)

  • Kai Dünser

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We are grateful to N Geldner, S Robert, K Schumacher, and T Ueda for sharing published material. The library of 360 bioactive small molecules was kindly provided by N Raikhel. We thank Maximilian Kaiser for helpful discussions about VAC1 and the BOKU-VIBT Imaging Center as well as the ZBSA-Life Imaging Center for access. We would like to thank Rosula Hinnenberg of the EM facility at the Faculty of Biology, University of Freiburg, for her assistance during the generation of EM data. The TEM (Hitachi HT7800) was funded by the DFG grant (project number 426849454) and is operated by the University of Freiburg, Faculty of Biology, as a partner unit within the Microscopy and Image Analysis Platform (MIAP) and the Life Imaging Center (LIC), Freiburg.

This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic (European Regional Development Fund-Project 'Plants as a tool for sustainable global development' No. CZ.02.1.01/0.0/0.0/16_019/0000827 to ON), The Austrian Academy of Sciences (ÖAW) (DOC fellowship to KD), Austrian science fund (FWF; P 33044 to JK-V), German Science fund (DFG; 470007283 and CIBSS – EXC-2189 to JK-V), and the European Research Council (ERC; 639478-AuxinER to JK-V).

Senior Editor

  1. Detlef Weigel, Max Planck Institute for Biology Tübingen, Germany

Reviewing Editor

  1. Heather E McFarlane, University of Toronto, Canada

Reviewer

  1. Takashi Ueda, National Institute for Basic Biology, Japan

Publication history

  1. Received: November 29, 2021
  2. Preprint posted: November 30, 2021 (view preprint)
  3. Accepted: May 31, 2022
  4. Version of Record published: June 10, 2022 (version 1)

Copyright

© 2022, Dünser et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 900
    Page views
  • 253
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Kai Dünser
  2. Maria Schöller
  3. Ann-Kathrin Rößling
  4. Christian Löfke
  5. Nannan Xiao
  6. Barbora Pařízková
  7. Stanislav Melnik
  8. Marta Rodriguez-Franco
  9. Eva Stöger
  10. Ondřej Novák
  11. Jürgen Kleine-Vehn
(2022)
Endocytic trafficking promotes vacuolar enlargements for fast cell expansion rates in plants
eLife 11:e75945.
https://doi.org/10.7554/eLife.75945

Further reading

    1. Cell Biology
    Lei Yuan et al.
    Research Article

    The phagocytic receptor CED-1 mediates apoptotic cell recognition by phagocytic cells, enabling cell corpse clearance in Caenorhabditis elegans. Whether appropriate levels of CED-1 are maintained for executing the engulfment function remains unknown. Here, we identified the C. elegans E3 ubiquitin ligase tripartite motif containing-21 (TRIM-21) as a component of the CED-1 pathway for apoptotic cell clearance. When the NPXY motif of CED-1 was bound to the adaptor protein CED-6 or the YXXL motif of CED-1 was phosphorylated by tyrosine kinase SRC-1 and subsequently bound to the adaptor protein NCK-1 containing the SH2 domain, TRIM-21 functioned in conjunction with UBC-21 to catalyze K48-linked poly-ubiquitination on CED-1, targeting it for proteasomal degradation. In the absence of TRIM-21, CED-1 accumulated post-translationally and drove cell corpse degradation defects, as evidenced by direct binding to VHA-10. These findings reveal a unique mechanism for the maintenance of appropriate levels of CED-1 to regulate apoptotic cell clearance.

    1. Cell Biology
    2. Genetics and Genomics
    Anthony J Veltri et al.
    Research Article Updated

    Key protein adapters couple translation to mRNA decay on specific classes of problematic mRNAs in eukaryotes. Slow decoding on non-optimal codons leads to codon-optimality-mediated decay (COMD) and prolonged arrest at stall sites leads to no-go decay (NGD). The identities of the decay factors underlying these processes and the mechanisms by which they respond to translational distress remain open areas of investigation. We use carefully designed reporter mRNAs to perform genetic screens and functional assays in Saccharomyces cerevisiae. We characterize the roles of Hel2, Syh1, and Smy2 in coordinating translational repression and mRNA decay on NGD reporter mRNAs, finding that Syh1 and, to a lesser extent its paralog Smy2, act in a distinct pathway from Hel2. This Syh1/Smy2-mediated pathway acts as a redundant, compensatory pathway to elicit NGD when Hel2-dependent NGD is impaired. Importantly, we observe that these NGD factors are not involved in the degradation of mRNAs enriched in non-optimal codons. Further, we establish that a key factor previously implicated in COMD, Not5, contributes modestly to the degradation of an NGD-targeted mRNA. Finally, we use ribosome profiling to reveal distinct ribosomal states associated with each reporter mRNA that readily rationalize the contributions of NGD and COMD factors to degradation of these reporters. Taken together, these results provide new insight into the role of Syh1 and Smy2 in NGD and into the ribosomal states that correlate with the activation of distinct pathways targeting mRNAs for degradation in yeast.