CSLD3 (KOJAK/RHD7, At3g03050) is a cellulose synthase-like D (CSLD) glycosyltransferase of the GT2 family. It has eight predicted transmembrane helices and a cytosolic catalytic domain with a conserved TED motif. Purified CSLD3 reconstituted into proteoliposomes uses UDP-glucose, not GDP-mannose, and makes beta-1,4-glucan. It assembles into large complexes with threefold-symmetric particles like those formed by CESA cellulose synthases, and its catalytic domain can replace that of CESA6 in vivo. A functional fluorescent CSLD3 fusion accumulates in a polarized apical plasma-membrane domain of growing root hairs, where CESAs are absent. There it makes cellulose-like beta-1,4-glucan needed for tip growth. The protein also passes through the Golgi and trans-Golgi network. Loss of CSLD3 lets root hairs form a bulge, but the hairs burst at the tip and fail to elongate, and their walls show disorganised cellulose and xyloglucan. CSLD3 shares partly redundant roles with CSLD2 (root hairs, female gametophyte) and with CSLD5 (whole-plant growth).
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: UniProt SubCell mapping from the curated "Golgi apparatus membrane" location. LOPIT proteomics independently puts CSLD3 in Golgi/TGN. Reason: CSLD3 is a multi-pass membrane protein, so Golgi membrane is the right level of precision for its Golgi presence, and two LOPIT studies support it (PMID:16618929, PMID:22923678). The functional evidence places catalytic activity at the apical plasma membrane of root hairs, not in the Golgi (PMID:21765420). As for CESAs, the Golgi is probably a trafficking compartment, so the term is kept as non-core. Supporting Evidence: UniProtKB:Q9M9M4 SUBCELLULAR LOCATION: Golgi apparatus membrane PMID:21765420 We also find that eYFP-CSLD3 proteins, but not CESA cellulose synthases, localize to a polarized plasma-membrane domain in root hairs. |
| GO:0005737 cytoplasm | IEA GO_REF:0000117 | TOO GENERAL | Summary: ARBA machine-learning location prediction for a broad compartment. Reason: True only in the sense that the Golgi and endomembranes lie within the cytoplasm. More specific locations kept in this review (Golgi membrane, trans-Golgi network, plasma membrane) already say where CSLD3 is, so the term adds nothing. |
| GO:0005768 endosome | HDA PMID:22923678 Putative glycosyltransferases and other plant Golgi apparatu... | KEEP AS NON CORE | Summary: LOPIT high-throughput localization. The endosome assignment probably reflects the plant TGN/early endosome, which LOPIT does not separate cleanly. Reason: In plants the TGN acts as the early endosome, and the same study also assigns CSLD3 to the TGN. This is a trafficking location for a protein whose activity is shown at the plasma membrane (PMID:21765420). It is kept, but as non-core. Supporting Evidence: PMID:22923678 we identified the subcellular localization of 1,110 proteins with high confidence |
| GO:0005783 endoplasmic reticulum | IDA PMID:11156607 KOJAK encodes a cellulose synthase-like protein required for... | KEEP AS NON CORE | Summary: Favery et al. 2001 saw a GFP-KOJAK fusion in the ER and from this argued for a non-cellulosic product. Reason: The observation is experimental. A multi-pass membrane protein will pass through the ER on its way out, so the term is not wrong. Later work with a functional, complementing eYFP-CSLD3 found the protein at the apical plasma membrane of root hairs (PMID:21765420), and purified CSLD3 makes beta-1,4-glucan (PMID:32169960). This overturns the paper's conclusion that an ER location implies a non-cellulosic product. ER is therefore not where the activity is exerted, and the term is kept only as non-core. Supporting Evidence: PMID:11156607 Subcellular localization of the KOJAK/AtCSLD3 protein using a GFP fusion shows that KOJAK/AtCSLD3 is located on the endoplasmic reticulum |
| GO:0005794 Golgi apparatus | HDA PMID:16618929 Mapping the Arabidopsis organelle proteome. | MODIFY | Summary: Organelle proteomics (LOPIT) assigns CSLD3 to the Golgi. Reason: The Golgi assignment is supported, but CSLD3 is an integral multi-pass membrane protein, so Golgi membrane (GO:0000139), which the gene already carries, is the precise term. As for CESAs, the Golgi is a trafficking compartment, not the site of activity. Proposed replacements: Golgi membrane Supporting Evidence: PMID:16618929 enabling confident and simultaneous localization of 527 proteins to the endoplasmic reticulum, Golgi apparatus, vacuolar membrane, plasma membrane, or mitochondria and plastids |
| GO:0005794 Golgi apparatus | HDA PMID:22923678 Putative glycosyltransferases and other plant Golgi apparatu... | MODIFY | Summary: LOPIT proteomics assigns CSLD3 to the Golgi apparatus. Reason: Same claim as the other Golgi apparatus rows. For a multi-pass membrane protein, Golgi membrane (GO:0000139, already on the gene) is the precise term. Proposed replacements: Golgi membrane Supporting Evidence: PMID:22923678 we identified the subcellular localization of 1,110 proteins with high confidence |
| GO:0005794 Golgi apparatus | ISM GO_REF:0000122 | MODIFY | Summary: AtSubP sequence-based location prediction. Reason: Matches the proteomic Golgi assignment. Refine to Golgi membrane (GO:0000139) because CSLD3 is a multi-pass membrane protein, as for the HDA rows. Proposed replacements: Golgi membrane |
| GO:0005802 trans-Golgi network | HDA PMID:22923678 Putative glycosyltransferases and other plant Golgi apparatu... | KEEP AS NON CORE | Summary: LOPIT proteomics places CSLD3 in the TGN. Reason: Plausible as a post-Golgi trafficking compartment on the way to the apical plasma membrane, but it is not where the synthase activity is shown to act (PMID:21765420). Supporting Evidence: PMID:22923678 we identified the subcellular localization of 1,110 proteins with high confidence |
| GO:0005886 plasma membrane | HDA PMID:14506206 Large-scale analysis of in vivo phosphorylated membrane prot... | ACCEPT | Summary: A CSLD3 phosphopeptide (Ser-755) was found in a plasma-membrane phosphoproteome. Reason: Agrees with direct imaging of a functional eYFP-CSLD3 in a polarized apical plasma-membrane domain of root hairs (PMID:21765420), where it acts as a (1->4)-beta-glucan synthase. The plasma membrane is where CSLD3 is active. Supporting Evidence: PMID:14506206 the identification of plasma membrane phosphoproteins of Arabidopsis PMID:21765420 We also find that eYFP-CSLD3 proteins, but not CESA cellulose synthases, localize to a polarized plasma-membrane domain in root hairs. |
| GO:0005886 plasma membrane | IBA GO_REF:0000033 | ACCEPT | Summary: PAINT node PTN001410221 (seeded by CESA7 and CESA6) is ancestral to the CESA and CSLD lineages. Reason: CSLD3 sits inside the clade descending from this node, and its own imaging confirms that it is active at the plasma membrane (PMID:21765420). The IBA and the gene-specific evidence agree. Supporting Evidence: file:interpro/panther/PTHR13301/PTHR13301-paint.tsv PTN001410221 GO:0005886 C IBD PMID:21765420 we provide evidence that CSLD3 represents a distinct (1β4)-Ξ²-glucan synthase activity in apical plasma membranes during tip growth in root-hair cells |
| GO:0009409 response to cold | IEA GO_REF:0000117 | MARK AS OVER ANNOTATED | Summary: ARBA machine-learning prediction, probably learned from the CSLD2/CSLD3 IEP rows. Reason: The only evidence is that CSLD3 expression rises at low temperature and that mutant root hairs rupture less at 5 Β°C (PMID:18768911, PMID:21279381). This shows the wall phenotype is temperature-sensitive. It does not show that CSLD3 takes part in a cellular response to cold. Supporting Evidence: PMID:18768911 the expression of both CSLD2 and CSLD3 is elevated at reduced temperatures |
| GO:0009409 response to cold | IEP PMID:18768911 Functional analysis of the cellulose synthase-like genes CSL... | MARK AS OVER ANNOTATED | Summary: Expression of CSLD3 is elevated at reduced temperature. Reason: An expression change alone (IEP) does not show participation in a cold response. The partial rescue of csld3 root hairs at low temperature is better explained by the physical behaviour of a weakened wall, possibly helped by compensating CSLD2 expression (PMID:21279381), than by a cold-signalling role. Supporting Evidence: PMID:18768911 the expression of both CSLD2 and CSLD3 is elevated at reduced temperatures PMID:21279381 CSLD2 could partially compensate for CSLD3 defects to prevent rupture at 5Β°C |
| GO:0009833 plant-type primary cell wall biogenesis | IBA GO_REF:0000033 | ACCEPT | Summary: PAINT node PTN001410221 (seeded by the primary-wall CESAs CESA1, CESA3 and CESA6) is ancestral to the CSLD lineage. Reason: Root hair walls are primary walls. CSLD3 makes beta-1,4-glucan for the growing root hair wall (PMID:21765420), its catalytic domain can stand in for that of CESA6 in primary-wall cellulose synthesis (PMID:32169960), and csld3 root-hair walls have disorganised cellulose and xyloglucan (PMID:21279381). The IBA is well supported by the gene's own data. Supporting Evidence: file:interpro/panther/PTHR13301/PTHR13301-paint.tsv PTN001410221 GO:0009833 P IBD PMID:21279381 These cell wall defects indicate that CSLD3 is required for the normal organization of both cellulose and xyloglucan in root hair cell walls. |
| GO:0012505 endomembrane system | IEA GO_REF:0000117 | TOO GENERAL | Summary: ARBA prediction of a very broad location. Reason: True, but it adds nothing next to the more specific Golgi membrane, trans-Golgi network and plasma membrane terms kept in this review. |
| GO:0016020 membrane | IEA GO_REF:0000002 | MODIFY | Summary: InterPro2GO mapping from the cellulose synthase domain (IPR005150). Reason: CSLD3 is an integral membrane protein, but "membrane" is too broad. The membrane where it is active is the plasma membrane (GO:0005886), which the gene already carries by HDA and IBA and which imaging supports (PMID:21765420). Proposed replacements: plasma membrane Supporting Evidence: PMID:21765420 We also find that eYFP-CSLD3 proteins, but not CESA cellulose synthases, localize to a polarized plasma-membrane domain in root hairs. |
| GO:0016760 cellulose synthase (UDP-forming) activity | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from the cellulose synthase catalytic domain. For CSLD3 this has been confirmed directly by biochemistry. Reason: GO:0016760 is defined as UDP-glucose + (1,4)-beta-D-glucan(n) = UDP + (1,4)-beta-D-glucan(n+1). Purified CSLD3 reconstituted in proteoliposomes used UDP-Glc, not GDP-Man, and made a product digested only by a beta-1,4-specific glucanase (PMID:32169960). The CSLD3 catalytic domain in CESA6 rescued the cesa6 mutant, but a TED-motif-dead version did not (PMID:32169960). The reverse chimera rescued csld3 root hairs (PMID:21765420). This is CSLD3's core molecular function, and it is established experimentally in Arabidopsis, not inferred. PAINT did not propagate cellulose synthase activity to the CSLD lineage (the GO:0016759 IBD sits on PTN000999295, which CSLD3 does not descend from). The activity here therefore rests on CSLD3's own biochemistry and should not be read as a family-wide transfer to other CSLD clades. Supporting Evidence: PMID:32169960 Taken together, these results strongly support the conclusion that CSLD3 utilizes UDP-Glc as a substrate and synthesizes Ξ²-1,4-glucan and not Ξ²-1,4-mannan. PMID:21765420 we provide evidence that CSLD3 represents a distinct (1β4)-Ξ²-glucan synthase activity in apical plasma membranes during tip growth in root-hair cells |
| GO:0030244 cellulose biosynthetic process | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping. It fits the experimentally shown beta-1,4-glucan synthase activity. Reason: CSLD3 polymerises beta-1,4-glucan from UDP-Glc (PMID:32169960). In root hairs it is needed for de novo synthesis of cellulose-like polysaccharides at the tip (PMID:21765420), and csld3 root-hair walls have disorganised cellulose (PMID:21279381). When its catalytic domain is placed in CESA6, it supports crystalline cellulose accumulation (PMID:32169960). CSLD3 catalyses the polymerisation step, so it takes part in the process. Whether the native CSLD3 polymer forms microfibrils like CESA cellulose is still open, but the polymer is chemically beta-1,4-glucan. Supporting Evidence: PMID:32169960 These studies further support the synthesis of Ξ²-1,4-glucan polymers that can integrate into crystalline cellulose. PMID:21765420 We show that cellulose-like polysaccharides are present in root-hair tips, and de novo synthesis of these polysaccharides is required for tip growth. |
| GO:0051753 mannan synthase activity | IDA PMID:21904114 Mannan synthase activity in the CSLD family. | REMOVE | Summary: Based on the Yin et al. 2011 / Verhertbruggen et al. 2011 tobacco microsome assays, in which GDP-Man mannosyltransferase activity was seen when CSLD2 and CSLD3 were co-expressed. Reason: The original data showed no mannan synthase activity for CSLD3 expressed on its own. Activity was seen only when it was co-expressed with CSLD2, in crude microsomes that already contain endogenous mannan synthase (PMID:21471331). Purified, reconstituted CSLD3 did not use GDP-Man and made beta-1,4-glucan from UDP-Glc, with CSLA9 as the positive control for mannan synthesis (PMID:32169960). The authors of that study noted the conflict with the microsome data and explained it as endogenous activity. Direct biochemistry contradicts CSLD3 having this activity, so the annotation is removed. The PTHR13301 family review reaches the same conclusion. Supporting Evidence: PMID:21471331 The same activity was detected when CSLD2 and CSLD3 were co-expressed but not when they were expressed separately. PMID:32169960 Taken together, these results strongly support the conclusion that CSLD3 utilizes UDP-Glc as a substrate and synthesizes Ξ²-1,4-glucan and not Ξ²-1,4-mannan. PMID:32169960 However, it should be noted that mannan synthesis in these studies was assessed in plant membranes overexpressing CSLD proteins that contained endogenous Ξ²-1,4-mannan activity |
| GO:0051753 mannan synthase activity | IEA GO_REF:0000117 | REMOVE | Summary: ARBA machine-learning prediction, which reflects the CSLD mannan IDA annotations. Reason: Purified CSLD3 does not use GDP-mannose and synthesises beta-1,4-glucan (PMID:32169960). The mannan assignment rests on co-expression microsome assays in which CSLD3 alone was inactive (PMID:21471331). Removed, consistent with the IDA row. Supporting Evidence: PMID:32169960 Taken together, these results strongly support the conclusion that CSLD3 utilizes UDP-Glc as a substrate and synthesizes Ξ²-1,4-glucan and not Ξ²-1,4-mannan. |
| GO:0048767 root hair elongation | IMP PMID:11156607 KOJAK encodes a cellulose synthase-like protein required for... | NEW | Summary: Proposed new annotation. kjk/csld3 mutants start root hairs normally, but the hairs burst at the tip and fail to elongate. CSLD3 supplies beta-1,4-glucan at the apical plasma membrane during tip growth. Reason: Participation test: the phenotype alone would only show necessity. Here CSLD3 also does part of the work: it makes the cellulose-like wall polysaccharide laid down at the growing tip (PMID:21765420), and de novo synthesis of that polysaccharide is needed for tip growth. Comparator check: the paralog CSLD2, which has the same role in the same cells, carries GO:0048767 (IMP, PMID:18768911). The term is absent for CSLD3, the founding mutant of this phenotype, so this looks like a gap rather than a deliberate convention. Two independent alleles show that elongation, not initiation, is affected (PMID:11156607, PMID:11402188). Supporting Evidence: PMID:11156607 Plants homozygous for kojak (kjk) mutations initiate root hairs that rupture at their tip soon after initiation. PMID:11402188 In the mutant, root hairs were initiated at the correct position and grew into a bulge, but their elongation was severely reduced. PMID:21765420 We show that cellulose-like polysaccharides are present in root-hair tips, and de novo synthesis of these polysaccharides is required for tip growth. |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: Should cellulose synthase (UDP-forming) activity be placed on a PAINT node covering the CSLD clade? CSLD3 is the only member characterised biochemically. Whether CSLD5 and the pollen CSLD1/CSLD4 share UDP-Glc specificity is untested, so a family-wide propagation would currently rest on one member.
Suggested experts: PAINT/PANTHER curators
Q: What makes the CSLD3 beta-1,4-glucan structurally different from CESA cellulose in tip walls, given that its catalytic domain can replace that of CESA6? Is the difference in complex architecture, in microtubule independence, or in co-polymerisation with other wall components?
Suggested experts: Plant cell wall biochemists (e.g. Nielsen lab)
Experiment: Express and purify CSLD2, CSLD5 and CSLD1 from yeast, reconstitute them into proteoliposomes as for CSLD3, and assay UDP-Glc and GDP-Man incorporation, with linkage-specific glucanase/mannanase product analysis.
Hypothesis: CSLD2 and CSLD5 share CSLD3's UDP-Glc-dependent beta-1,4-glucan synthase activity.
Experiment: Use native-promoter tagged CSLD3 with CSLD2 co-immunoprecipitation and single-particle imaging of root-hair plasma membranes to find whether CSLD2 or CESAs co-assemble with CSLD3.
Hypothesis: Native CSLD3 complexes at the root hair tip are homo-oligomeric.
Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)