Cellulose is one of the most abundant biopolymers on the planet, and this para-crystalline polysaccharide is a major component of plant cell wall extracellular matrices. Cellulose is synthesized at the plasma membrane of plant cells (Paredez et al. 2006) by Cellulose Synthase Complexes (CSCs), which are comprised of multiple non-redundant Cellulose Synthase A (CESA) subunits that serve as the CSC catalytic subunits (Purushotham et al. 2016;Cho et al. 2017). The CSC also contains accessory subunits, such as the KORRI-GAN endoglucanase (Vain et al. 2014), Cellulose Synthase Interactive 1 (Gu et al. 2010;Li et al. 2012), Companion of Cellulose synthase (CC) proteins (Endler et al. 2015), and the glycosylphosphatidylinositol (GPI)-anchored protein COBRA (Borner et al. 2003;Li et al. 2013;Liu et al. 2013). Together, these subunits form a massive molecular machine that extrudes newly synthesized cellulose into the apoplast (Purushotham et al. 2020). Live-cell imaging has revealed that CSC localization is dynamic and complex. Fluorescently labeled CSC subunits reside in the Golgi apparatus, small vesicular compartments known as Microtubule Associated Cellulose Synthase Compartments (MASCs) or Small CESA Compartments (SmaCCs) (Crowell et al. 2009;Gutierrez et al. 2009), and as motile particles at the plasma membrane that move with constant velocities of approximately 250 nm/min (Paredez et al. 2006). Plasma membrane localized puncta are thought to represent actively synthesizing CSCs, and the rate of CSC motility has been correlated to cellulose biosynthetic output in numerous inhibitor and genetic studies (Paredez et al. 2006;Debolt et al. 2007;Paredez et al. 2008;Gu et al. 2010). Despite these insights, the mechanistic factors that govern CSC motility and therefore activity at the plasma membrane are not fully elucidated.
Sphingolipids are a unique group of lipids that play major structural roles in plant membrane architecture as well as critical roles in programmed cell death, membrane trafficking, plant pathogen interactions, protein anchoring to the plasma membrane, and developmental processes (Ternes et al. 2011;Gronnier et al. 2016). Sphingolipids consist of a fatty acid (FA) group with an acyl chain ranging from 16 to 26 carbons, an 18-carbon sphingosine base, commonly referred to as a Long Chain Base (LCB) that can be di-(d18) or tri-(t18) hydroxylated, and a structurally variable headgroup. In Arabidopsis, LCBs and FA-CoAs are used as substrates for the ceramide synthase enzymes LONGEVITY ASSURANCE GENE ONE HOMOLOG 1 (LOH1), LOH2, and LOH3 to form ceramides (Markham et al. 2011;Ternes et al. 2011). These enzymes have distinct substrate specificities for different forms of LCBs and FA-CoAs (Luttgeharm et al. 2016). Ceramide FA moieties can be hydroxylated to form hydroxyceramides (hCers), and these sphingolipids often undergo further structural modifications, such as the attachment of glucose at the carbon 1 position of the LCB by GLUCO-SYLCERAMIDE SYNTHASE (GCS) (Msanne et al. 2015) to form glucosylceramides (GlcCers). Alternatively, hCers can be used as substrates for the synthesis of Glycosylinositolphosphoceramides (GIPCs) (Rennie et al. 2014;Fang et al. 2016). The elaborate headgroup of GIPCs contains a phosphate attached to the LCB, followed by an inositol residue that is attached by INOSITOL PHOSPHORYLCERAMIDE SYN-THASE (Wang et al. 2008). INOSITOL PHOSPHORYL-CERAMIDE GLUCURONOSYLTRANSFERASE 1 (IPUT1) uses UDP-Glucuronic acid as a sugar donor to attach glucuronic acid to this inositol moiety to form the core GIPC structure (Tartaglio et al. 2017), which can be further elaborated by the addition of hexose and pentose sugars (Fang et al. 2016;Ishikawa et al. 2018).
Impaired GIPC biosynthesis causes poor growth and reduced cellulose content in Arabidopsis seedlings (Fang et al. 2016), but the mechanistic basis of this observation is unclear. Based on previous work indicating that sphingolipids participate in membrane trafficking and organization, it is possible that GIPCs or other sphingolipids are required for proper transport of the CSC from the Golgi or vesicles to the plasma membrane. It is also unclear whether synthesis of other sphingolipid structural classes is required for proper cellulose deposition, and this issue is exacerbated by the observation that Arabidopsis mutants which are specifically compromised in glycosylated sphingolipid biosynthesis are gametophytically or seedling lethal (Rennie et al. 2014;Msanne et al. 2015;Tartaglio et al. 2017).
Prior work has highlighted that various molecules based on deoxynojirimycin scaffolds can impact sphingolipid biosynthesis by inhibiting the synthesis or catabolism of these important lipid species (Ashe et al. 2011;Rugen et al. 2018). Here, we assayed a group of nojirimycin analogs against Arabidopsis thaliana to identify compounds that caused root growth inhibition and root swelling, which are hallmarks of cell wall biosynthesis inhibition. From this chemical screen, N-dodecyl deoxynojirimycin (ND-DNJ) was identified as a robust inhibitor of root growth. Subsequent experiments revealed that ND-DNJ inhibits the biosynthesis of glycosphingolipids, and also inhibits the deposition of crystalline cellulose. These effects were recapitulated in an Arabidopsis mutant that is compromised in GlcCer biosynthesis. Live-cell imaging of CSCs in Arabidopsis seedlings treated with ND-DNJ also revealed that this compound caused reduced CSC motility, suggesting that GlcCers are required for CSC activity and motility.
Prior work has indicated that a select subset of nojirimycinbased compounds significantly impact plant growth and development (Rugen et al. 2018). Based on our interest in identifying inhibitors of plant cell wall glycosyltransferases (Xia et al. 2014;Villalobos et al. 2015), we screened a small panel of commercially available nojirimycin analogs (Supplemental Fig. 1A) that was largely orthogonal to a previous study (Rugen et al. 2018) for their ability to induce cell wall biosynthetic defects, including reduced root elongation and root swelling. To evaluate the effects of this inhibitor panel, Arabidopsis seedlings were germinated on MS medium containing 100 μM test compound, grown for 7 days as described in Materials and Methods, and their root lengths were quantified (Supplemental Fig. 1B). These experiments revealed that N-dodecyl-deoxynojirimycin (ND-DNJ) (Fig. 1A) robustly inhibited Arabidopsis root growth, while the majority of additional compounds in the screen were only moderately effective or completely inactive in this assay.
To further investigate the phenotypic defects elicited by ND-DNJ, we performed a series of detailed growth assays. Consistent with the initial screening assays, 100 μM ND-DNJ treatment inhibited Arabidopsis primary root growth by nearly 80% (Fig. 1B and C) and inhibited Arabidopsis darkgrown hypocotyl elongation by 50% (Fig. 1B and D). Further detailed microscopy analysis and quantification revealed that 100 μM ND-DNJ treatment caused Arabidopsis primary roots to swell by 30% compared to untreated or solvent controls (Supplemental Fig. 2A). A dose-response curve was constructed to measure the effect of increasing ND-DNJ concentrations on primary root elongation (Supplemental Fig. 2B). These assays revealed that ND-DNJ inhibited primary root elongation in a dose-dependent manner with an apparent IC 50 of 21 μM. Overall, these results indicate that ND-DNJ is an effective inhibitor of primary root growth, and potentially suggest that this inhibitor causes defects in cell expansion that are linked to inhibition of cell wall polysaccharide deposition.
Molecules that are structurally similar to ND-DNJ inhibit glucocerebrosidases in animal systems, impact sphingolipid degradation, and serve as an effective drug treatment for Gaucher's disease as substrate assisted chaperones (Stirnemann et al. 2017). A ceramide analog, N-[2-hydroxy-1-(4-morpholinylmethyl)-2-phenyl ethyl]-decanamide (PDMP) has also been widely used as a GlcCer biosynthesis inhibitor to study vesicular trafficking defects associated with GlcCer depletion (Melser et al. 2010), and a recent screen of nojirimycin-based analogs, which did not include ND-DNJ, also suggested that nojirimycin-based molecules are GlcCer biosynthesis inhibitors (Rugen et al. 2018). Importantly, N-5-(adamantane-1-yl-ethoxy) pentyl-L-ido-deoxynojirimycin (Lido-AEP-DNJ) produced similar phenotypic defects compared to ND-DNJ (Rugen et al. 2018), and we confirmed in our own experiments that PDMP inhibits root and hypocotyl elongation to a similar degree as ND-DNJ (Supplemental Fig. 3). Therefore, we postulated that ND-DNJ may inhibit GlcCer biosynthesis in a similar manner to these structurally related molecules.
To test this hypothesis, Arabidopsis seedlings grown for 7 days in the presence of 100 μM ND-DNJ or 50 μM PDMP were lyophilized and subjected to established sphingolipid extraction procedures as described in Materials and Methods. Untreated plants and solvent control seedlings grown in either 0.1% methanol or 0.1% DMSO for ND-DNJ and PDMP, respectively, were processed in the same fashion. The resulting sphingolipid extracts were subjected to targeted lipidomic analysis to quantify the abundance of various sphingolipid metabolites, including free long chain bases (LCBs), GlcCers, GIPCs, hydroxyceramides (hCers), and ceramides (Cers). These experiments revealed that ND-DNJ and PDMP treatment elicited clear impacts on sphingolipid biosynthesis (Fig. 2). Arabidopsis seedlings treated with ND-DNJ exhibited a 30% reduction in total GlcCer content compared to untreated controls, with decreases primarily reflected in GlcCers containing d18:1 and d18:2 LCBs (Fig. 2D and E). In contrast, total GIPC content was minimally impacted by ND-DNJ treatment, with only GIPC species containing d18:1 LCBs exhibiting a significant reduction (Fig. 2C and E). ND-DNJ treatment also resulted in nearly a 10-fold increase in total free LCB content (Fig. 3A) with major increases in free d18:0, t18:0, and t18:1 LCBs. Notably, many phosphorylated LCBs were also elevated with d18:1-P, t18:0-P, and t18:1-P being the most prominent (Fig. 3C). Upon further analysis of d18:1 and t18:1 LCBcontaining GlcCers, the most prominent GlcCer species for each LCB, we found that the abundance of d18:1 LCBcontaining GlcCers containing C16 acyl groups was reduced by 40% (Supplemental Fig. 4) in ND-DNJ treated seedlings. In contrast, t18:1 LCB-containing GlcCers with C16 acyl groups increased in abundance by 50%, suggesting that d18 GlcCers are preferentially impacted by ND-DNJ treatment. We also noted that the concentrations of the highly abundant t18:1 GlcCers with acyl groups ranging from 22 to 26 carbons were minimally impacted by ND-DNJ, but concentrations of d18:1 GlcCers containing longer chain acyl groups were reduced by nearly 5-fold. This analysis supports the hypothesis that ND-DNJ reduced GlcCer content in Arabidopsis.
Consistent with its ascribed function as a GCS inhibitor (Inokuchi et al. 1989;Barbour et al. 1992;Fenderson et al. 1992;Rugen et al. 2018), PDMP treatment caused a 40% reduction in total GlcCer content, with substantial decreases in both d18 (46%-65%) and t18 (30%-40%) LCB-containing GlcCers (Fig. 2D and E). PDMP caused only a 1.8-fold increase in total LCB species compared to the 9.3-fold increase elicited by ND-DNJ treatment (Fig. 3B). Interestingly, GIPC profiles were also impacted by PDMP treatment. Total GIPC content was reduced by 39% in PDMP-treated seedlings, reflected primarily by a reduction in t18:0 and t18:1 LCB species (Fig. 2C and E). Overall, these results suggest that ND-DNJ and PDMP have complex effects on Arabidopsis sphingolipid content and that the common outcome of treatment with these inhibitors is a substantial reduction in GlcCer species.
While sphingolipids are required for numerous physiological processes in plants, it is important to note that inhibition of GIPC glycosylation via genetic disruption of Arabidopsis GIPC MANNOSYL-TRANSFERASE1 caused substantial reduction in crystalline cellulose content (Fang et al. 2016). However, the impact of GlcCer metabolism on cellulose deposition has not been previously examined. Therefore, we analyzed matrix polysaccharide composition and crystalline cellulose content (Villalobos et al. 2015) of 7-day-old seedlings treated with 100 μM ND-DNJ or 50 μM PDMP. Six commonly occurring cell wall monosaccharides were assayed by alditol acetate analysis, and these experiments revealed that arabinose content increased nearly 2-fold in ND-DNJ treated seedlings while the other five cell wall matrixassociated monosaccharides remain relatively unchanged (Fig. 4A). A similar elevation of arabinose content was not observed under PDMP treatment (Fig. 4B). Intriguingly, total matrix polysaccharide glucose contents of ND-DNJ and PDMP-treated seedlings were nearly five times greater compared to untreated and solvent controls (Fig. 4E and F). This observation prompted further analysis of the matrix polysaccharide glucose content in ND-DNJ and PDMP treated seedlings. Using 10 mg of the remaining AIR sample, we removed starch using porcine amylase and pullulanase overnight. Enzymes were removed with solubilized glucose, and the remaining tissue was subjected to alditol acetate assay. Glucose content in ND-DNJ and PDMP-treated seedlings was still elevated compared to untreated and solvent controls (Supplemental Fig. 5), potentially suggesting that this increase in matrix polysaccharide glucose content is the result of an additional glucan, such as wound-induced callose or non-crystalline cellulose. Crystalline cellulose quantification analysis was also investigated via Updegraff analysis (Updegraff 1969), and these experiments demonstrated that seedlings germinated under ND-DNJ or PDMP treatment displayed at 25%-35% reduction in crystalline cellulose content (Fig. 4C and D). Overall, these results indicate that ND-DNJ and PDMP treatment cause a significant reduction in crystalline cellulose content with a concomitant increase in matrix polysaccharide glucose content, suggesting that GlcCer biosynthesis is necessary for the proper deposition of crystalline cellulose.
GlcCers are produced via the action of GCS, and only one GCS enzyme has been identified in the Arabidopsis genome (Msanne et al. 2015). gcs-1 loss-of-function mutants germinate but exhibit seedling growth arrest after several days under light-grown conditions. To investigate whether lossof-function mutations in the GCS gene cause similar effects compared to ND-DNJ or PDMP treatment, we grew the previously described gcs-1 mutants on MS medium in the light and first measured root elongation. As demonstrated in Fig. 5B, gcs-1 mutant root lengths were reduced by 90% compared to wild-type Col-0 controls. Thus this mutant exhibits Next, we investigated the cell wall composition of gcs-1 mutant plants compared to wild-type Col-0 controls. Alcohol Insoluble Residue (AIR) from 7-day-old light grown seedlings was prepared as described in Materials and Methods, and the resulting AIR samples were subjected to alditol acetate monosaccharide analysis as well as crystalline cellulose quantification as described above. These experiments revealed that gcs-1 mutant plants exhibited 2-fold higher matrix polysaccharide glucose contents, similar to observations of ND-DNJ and PDMP treated plants. The remaining matrix polysaccharide-associated monosaccharides were unchanged in gcs-1 plants compared to wild-type Col-0 controls (Fig. 5C). Crystalline cellulose content of gcs-1 mutants was also measured (Fig. 5D), revealing that the gcs-1 mutant exhibited a 42% reduction in crystalline cellulose content. Overall, these results suggest that genetic disruption of Arabidopsis GCS results in similar phenotypic and cell wall chemistry defects compared to ND-DNJ and PDMP treatment, further supporting the hypothesis that disruption of GlcCer biosynthesis causes the inhibition of cellulose deposition.
Cellulose is synthesized by CSCs in the plasma membrane, which can be visualized in living cells by examining the dynamics of fluorescent protein-CSC subunit fusions (Paredez et al. 2006;Debolt et al. 2007;Gutierrez et al. 2009). This imaging system is a useful tool to investigate the mechanism of inhibitors that impact cellulose biosynthesis (Paredez et al. 2006;Debolt et al. 2007;Xia et al. 2014). To examine the impact of PDMP and ND-DNJ treatment on CSCs, we treated dark-grown GFP-CESA3-expressing transgenic seedlings with 100 μM ND-DNJ or 50 μM PDMP for 1 h and examined CSC speed distributions in hypocotyl epidermal cells of these seedlings via spinning disc confocal microscopy and livecell cinematography as described in Materials and Methods. Untreated GFP-CESA3 seedlings, or seedlings treated with 0.1% solvent served as controls. As shown in Fig. 6A, singleframe images of epidermal cells treated with PDMP or ND-DNJ did not reveal drastic differences in CSC distribution throughout the plasma membrane, suggesting that GlcCer biosynthesis inhibition does not significantly impact the positional distribution or overall number of CSCs at the plasma membrane. However, time-averaged projections of CSC trajectories over a 5-min time-lapse revealed that CSCs in ND-DNJ and PDMP-treated seedlings showed shorter motility trajectories than corresponding controls. Indeed, Kymograph analysis of individual CSC particles (Fig. 6B) revealed that untreated plants exhibited linear constant velocities that are typically observed of motile CSCs, while the CSCs of ND-DNJ and PDMP-treated samples exhibited nearly vertical kymographs, suggesting that CSC velocities were reduced under these treatments. We then measured the velocity distribution of CSC particles from 6-10 cells from at least 3 independent seedlings per treatment group (Fig. 6C) and found that untreated seedlings exhibited an average speed of 264 nm/min, consistent with previous observations (Paredez et al. 2006;Gutierrez et al. 2009). While this average speed was not significantly impacted by solvent control treatment, average CSC speeds in ND-DNJ and PDMP-treated seedlings were reduced to 107 and 147 nm/min, respectively. We additionally generated multiple gcs-1 mutant lines expressing fluorescent protein-CESA markers, but unfortunately, gcs-1 homozygous mutants did not germinate under dark grown conditions, precluding further analysis of CSC dynamics in this genetic background. Overall, these observations indicate that ND-DNJ and PDMP treatment cause decreased CSC motility, suggesting that inhibition of GlcCer biosynthesis by these compounds causes the inhibition of crystalline cellulose deposition by reducing CSC motility and activity.
Sphingolipid biosynthesis impacts numerous aspects of plant growth and development, and previous work has suggested a functional relationship between the synthesis of GIPCs and cellulose deposition (Fang et al. 2016), but the mechanistic details of this relationship are unclear and the impact of other sphingolipids on cellulose deposition has not been investigated. Here, we demonstrate that ND-DNJ in an inhibitor of glycosylated sphingolipid biosynthesis (Fig. 2), which primarily impacts synthesis of GlcCers and LCBs. Treatment of Arabidopsis seedlings with this compound caused significant phenotypic abnormalities that are commonly associated with cell wall biosynthesis inhibition (Fig. 1; Supplemental Fig. 2). Treatment of Arabidopsis seedlings with ND-DNJ or the previously described GlcCer biosynthesis inhibitor PDMP (Inokuchi et al. 1989;Barbour et al. 1992;Fenderson et al. 1992;Rugen et al. 2018) also caused defects in cell wall polysaccharide deposition, including reduced crystalline cellulose content and increased cell wall matrix glucose content (Fig. 4). Genetic disruption of GCS recapitulated these biochemical defects (Fig. 5), strongly suggesting that GlcCers are required for normal crystalline cellulose deposition in Arabidopsis. Finally, the mechanism of cellulose biosynthesis inhibition elicited by ND-DNJ and PDMP treatment was investigated by examining the dynamics of CSCs via live-cell imaging, and these experiments revealed that CSC motility, and likely activity, was impaired upon treatment with ND-DNJ or PDMP (Fig. 6). Overall, these results suggest that reduced GlcCer content negatively impacts CSC motility and the deposition of crystalline cellulose in the plant cell walls of Arabidopsis.
Both ND-DNJ and PDMP have substantial impacts on GlcCer metabolism, but differences between the detailed sphingolipid species profiles in plants treated with these inhibitors indicate that these molecules may have unique mechanisms of action. PDMP is a well-established inhibitor of GlcCer biosynthesis in metazoan systems (Inokuchi et al. 1989;Barbour et al. 1992;Fenderson et al. 1992), and this compound has been used in multiple plant species to investigate the impact of GlcCer depletion on various cellular and physiological processes (Melser et al. 2010;Krüger et al. 2013;Rugen et al. 2018). Here, we have demonstrated that PDMP reduces total GlcCer content and impacts both d18 and t18 LCBcontaining species to a similar degree, consistent with the idea that this compound inhibits GCS in Arabidopsis. The majority of other sphingolipid species were not substantially impacted by PDMP treatment. While ND-DNJ elicited similar effects on GlcCer contet, it also caused a nearly 10-fold increase in total free LCB content, and the majority of these more abundant species are t18 LCB-related metabolites. Additionally, ND-DNJ caused a 2-fold increase in matrix polysaccharide arabinose content that was not observed upon PDMP treatment. Although this finding is outside the scope of this manuscript, it may suggest that accumulation of LCBs, which can serve as cell death signals, may lead to increased accumulation of arabinose containing polymers, such as pectic arabinan or arabinogalactan protein glycans, and these observations will be the subject of future investigation. Overall, PDMP and ND-DNJ have very similar impacts on cellulose biosynthesis, but different matrix polysaccharide and sphingolipid profiles, suggesting that ND-DNJ and PDMP both reduce total GlcCer content, but potentially through different mechanisms.
We used chemical genetic approaches and affinity purification techniques to attempt the identification of the ND-DNJ target, but despite our best efforts, these approaches were not successful. Thus, the molecular target of ND-DNJ remains to be identified, which is a limitation of this study and will be the focus of future work. However, we suggest that the observed sphingolipid metabolism defects could be explained by ND-DNJ activating the de-acylation or degradation of GlcCers. In this model, GlcCer specific ceramidases or de-glucosylation of GlcCers would occur through the action of glucosylceramidases. The first of these activities has not been described to our knowledge in plants, but a small family of four glucosylceramidase enzymes were recently described in the Arabidopsis genome (Dai et al. 2020). Recombinant Arabidopsis GLUCOSYLCERAMIDASE3 (GCD3) preferentially hydrolyzed glucose moieties of d18 and t18 GlcCers with acyl chains between 18-20 carbons (Dai et al. 2020), which could explain the preferential inhibition of d18 LCBcontaining GlcCer synthesis upon ND-DNJ treatment. Lossof-function mutants in GCD3 did not exhibit observable phenotypes and did not recapitulate the phenotypes exhibited upon ND-DNJ treatment, but it is important to note that multiple isoforms of GCD were identified, and these genes could be partially redundant or may need to be overexpressed to phenocopy the impacts of ND-DNJ. Importantly, Arabidopsis GCD enzymes are the plant homologs of mammalian glucocerebrosidases, and these enzymes are known to be allosterically activated or inhibited by a broad range of imino sugarbased small molecules (Kopytova et al. 2021), suggesting that GCD allosteric activation by ND-DNJ is a plausible mechanism to explain the associated impacts of this molecule. While the mechanism of ND-DNJ remains unknown and will be the subject of future investigations, it is important to note that PDMP treatment and the gcs-1 mutant exhibited reduced crystalline cellulose content and PDMP elicits similar CSC motility defects compared to ND-DNJ, thus supporting the hypothesis that GlcCer content is important for cellulose deposition.
The inhibition of GlcCer metabolism strikingly reduced speed distribution of CSCs treated with ND-DNJ or PDMP, suggesting that GlcCer content impacts CSC motility and activity. We postulate that this effect could arise from one of at least two possible models. First, GlcCers and other sphingolipids are enriched in membrane nanodomains within the plasma membrane that scaffold protein complexes, and prior work has demonstrated that d18 LCB-containing sphingolipids are particularly enriched in these membrane nanodomains (Borner et al. 2005;Carmona-Salazar et al. 2021) along with cellulose synthase activity (Bessueille et al. 2009). Therefore, a plausible explanation for the reduced speed distribution of CSCs upon GlcCer depletion is that the CSC is encased in a lipid nanodomain of sphingolipids within the plasma membrane, and that disruption of this nanodomain causes the reduced velocity of the CSC. GPI-anchored proteins are often preferentially enriched in membrane nanodomains (Borner et al. 2003;Borner et al. 2005), and the GPI-anchored protein COBRA (COB) is genetically required for cellulose biosynthesis (Roudier et al. 2005). Previous work suggest that COB is specifically required for the deposition of crystalline cellulose (Liu et al. 2013), suggesting that the disruption of GlcCer metabolism could disassociate COB from the CSC and cause the CSC to deposit cellulose with reduced crystallinity. Overall, the results presented here indicate that reducing Arabidopsis GlcCer content via chemical and genetic approaches compromises CSC activity, motility, and the deposition of crystalline cellulose, thus suggesting that these sphingolipids may actuate aspects of plant growth and development by regulating cellulose deposition.
A. thaliana Columbia (Col-0) seeds were sterilized in seed sterilization solution (30% [v/v] bleach, 0.1% SDS [w/v]) for 20 min at 25 • C. This solution was removed, and seeds were washed 5 times in sterile water followed by incubation at 4 • C for 48 h. Seeds were germinated on MS media (1/2 X Murashige and Skoog salts, 10 mM MES-KOH pH 5.7, 1% Sucrose [w/v], and 1% phytoagar [w/v]). N-Dodecyl-Deoxynojirymycin (ND-DNJ) and all other compounds described in this study were purchased from Toronto Research Chemicals (TRC) (Ontario, Canada) and added to MS agar media at specified concentrations after the media was autoclaved and cooled to below 50 • C. Seedlings were grown under long day conditions (16 h light, 8 h dark) at 22 • C vertically for 7 days. Seedling roots were straightened and scanned with a benchtop scanner, root lengths were quantified using ImageJ. For assays involving dark-grown seedlings, MS media lacking sucrose was supplemented with ND-DNJ or PDMP at the indicated concentrations. Seeds were sterilized as described above, plated onto MS media lacking sucrose and incubated in the light for 1 h. Plates were subsequently wrapped in aluminum foil and grown under long day conditions at 22 • C as described above for 5 days. Etiolated hypocotyl lengths were quantified using ImageJ.
The gcs-1 (SK-2634) mutant allele was obtained from the Arabidopsis Biological Resource Center. Genomic DNA extraction and PCR genotyping was performed as previously described (Smith et al. 2018). The gcs-1 mutant allele genotyping was conducted with the following primer set: SK-2634 LP: CCTCATCCTTCGGTCAATATG, SK-2634 RP: GGGACTCTTGGGAGTTATTGC, and SK T-DNA LB: ATACGACGGATCGTAATTTGTCG.
Alcohol Insoluble Residue (AIR) was prepared from 7-day-old light-grown Arabidopsis seedlings, and this material was used to perform alditol acetate analysis of cell wall matrix polysaccharides as previously described (Villalobos et al. 2015). Crystalline cellulose content was quantified by the Updegraff assay (Updegraff 1969) with modifications. Five hundred microliters of Updegraff reagent ([8:1:2] [v:v:v] Acetic acid: Nitric acid: water) was added to 10 mg of AIR material, and the samples were heated at 100 • C for 30 min. Samples were centrifuged at 14,000 × g for 15 min at 25 • C, and the supernatant was removed. The pellet was washed with 750 μL of water, followed by centrifugation under the same conditions. The supernatant was removed, and the resulting pellet was air dried at 80 • C. Samples were resuspended in 175 μL of 72% [v/v] H 2 SO 4 and incubated at 22 • C for 30 min. Samples were vortexed and incubated at 22 • C for an additional 15 min, followed by dilution with 825 μL of water, and 200 μL of each sample was transferred to a 1.5 mL tube. Four hundred microliters of freshly prepared Anthrone reagent (2% [w/v] Anthrone in concentrated H 2 SO 4 ) was added to each sample and heated to 80 • C for 10 min, then cooled to 22 • C. Seventy-five microliters of each sample was arrayed into a 96-well flat bottom plate, and sample absorbance was measured at 625 nm in a SpectraMax M5 spectrophotometer. Glucose content in each sample was calculated based on linear regression from glucose standards.
Arabidopsis seedlings were germinated in 1/2 X MS + 1% Sucrose media without supplementation or containing 0.1% Methanol, 100 μM ND-DNJ, 0.1% DMSO, or 50 μM PDMP. Seedlings were grown for 7 days under long day conditions in the light as described above, and then transferred to a 15 mL tube, flash frozen in liquid nitrogen, and lyophilized using a Labconco freeze dry system. Sphingolipid extraction and targeted lipidomic analysis were performed using LC-MS/MS as previously described (Markham and Jaworski 2007;Kimberlin et al. 2013).
A transgenic Arabidopsis line expressing GFP-CESA3 under its native promoter was previously described (Gutierrez et al. 2009). This line was grown on MS agar without sucrose for 4 days in the dark, and etiolated seedlings were mounted in 200 μL of water, 0.1% MeOH, 100 μM ND-DNJ, 0.1% DMSO, or 50 μM PDMP for 1 h prior to imaging. Seedlings were mounted as previously described (Paredez et al. 2006) and imaged on a Leica DMI-8 confocal microscope featuring a CSU-W1 spinning disk head (Yokogawa), a 100× 1.47NA oil immersion objective (Leica), and an iXon Life EMCCD camera (Andor Technology) controlled by VisiView software (Visitron Systems). GFP fluorescence was excited with a 488 nm laser and emission was recorded with a 525 nm filter. CSC motility assays were performed by recording 5-min time-lapse movies in 10 s increments.
All time-lapse recordings were processed using identical processing conditions in Fiji (https://imagej.net/Fiji). Movies were corrected for drift using rigid body alignment with the StackReg plugin (Thevenaz et al. 1998). Time-lapse images were further processed by performing a 4-frame walking average, background subtraction of 50 pixels, and then contrast enhancement. Velocities were analyzed using the "kymograph evaluation" plug-in of Fiji as previously described (Verbančič et al. 2021). Approximately 100 velocity measurements were collected from each cell, and cells from at least 6-10 seedlings were used to produce CSC speed distributions.
Downloaded from https://academic.oup.com/glycob/article/34/6/cwae035/7660701 by University of Georgia Libraries user on 05 June 2025