Interaction with COPII Member SAR1 Is Critical for the Delivery of Arabidopsis Xyloglucan Xylosyltransferases XXT2 and XXT5 to the Golgi Apparatus

Abstract

Transport of Golgi-localized proteins from the ER is mediated by the coat protein complex II (COPII) and its members, COPII inner coat subunit Sec24 and Secretion-associated Ras-related GTPase 1 (Sar1). Sar1 and Sec24 recognize cytosolic N-termini of glycosyltransferases (GTs) that contain peptide signals required for incorporation into COPII-coated vesicles. Xyloglucan Xylosyltransferases (XXTs) are required for xyloglucan (XyGs) biosynthesis and must be transported to the Golgi for proper function. In this study, we demonstrated that XXTs interact with AtSar1 in the COPII complex but not with AtSec24, which was previously reported to be the main recruiter of cargo proteins into COPII-coated vesicles. The mutation of the arginine to glutamine residues of di-arginine motifs in the N-termini of XXTs caused protein mislocalization and significantly reduced the strength of the interaction with AtSar1. These mutations caused 90% of XXTs to either remain in the ER or localize to small non-Golgi compartments. In turn, such mislocalization significantly suppressed the recovery of XyGs biosynthesis in Arabidopsis thaliana (Arabidopsis) mutants (xxt1xxt2 and xxt3xxt4xxt5), failing to restore their root phenotypes to normal. Our results demonstrate the interaction between cargo and AtSar1, highlighting the critical role of di-arginine motifs in this interaction. These results provide new insights into the mechanism of ER-to-Golgi delivery of plant GTs, which significantly advances our understanding of polysaccharide biosynthesis in the Golgi and the enzymes responsible for it.

1. Introduction

XyGs are a major hemicellulose and a significant component of the plant primary cell wall. XyGs are highly branched polysaccharides with a β-1,4-linked glucan backbone, branched with side chains composed of diverse monosaccharides [1,2,3,4,5]. XXTs transfer xylosyl residues to the glucan backbone as the first step of the XyG side chain during XyG biosynthesis in the Golgi [6,7,8]. Then, xylosyl residues are further substituted with sugars, such as galactose, fucose, arabinose, or glucuronic acid, generating diverse side-chain patterns [9,10,11,12,13,14]. In Arabidopsis, XyGs consists predominantly of XXXG, XLXG, XXFG, and XLFG subunits [15,16]. XXTs are a subset of GTs, a large family of enzymes that glycosylate diverse molecules in the Golgi and ER [8,17]. XXTs are type II membrane proteins and share a basic structure with most GTs, including an N-terminal cytosolic tail, a transmembrane domain (TMD), a stem region, and a C-terminal catalytic domain [15,18].

The N-terminal cytosolic tails and TMD of GTs have been shown to play a key role in their trafficking from the ER to the Golgi in eukaryotic cells [16,19,20,21,22,23,24], where specific amino acid motifs are critical for their Golgi localization. For example, arginine residues in the cytosolic tail of N-glycosylating GTs Arabidopsis β-1,2 xylosyltransferase (AtXylT) and Nicotiana tabacum (N. tabacum) N-acetylglucosaminyltransferase I (NtGnTI) are required for their recruitment to the Golgi [20]. In addition, it was demonstrated that the specific motifs in the cytosolic tails of UDP-GalNAc polypeptide N-acetylgalactosaminyltransferases (GalNAc-T1 and GalNAc-T2) and N-acetylglucosamine-1-phosphotransferase (PT) α/β-subunits are essential for their Golgi localization [19,21,22,24]. In yeast cells, sorting signal peptides at the protein N- or C-termini directly or indirectly interact with COPII components in the coat vesicles. Sar1 proteins, part of the COPII complex, directly interact with cargo proteins [25,26], including GTs [19,27], and their interactions can affect the assembly of COPII coat vesicles. Cytosolic tails of cargo GTs directly interact with Sar1 and Sec23p to promote COPII trafficking, exemplified by in vitro binding of yeast Sar1 to di-arginine motif tails from GalNAc-T and β-1,3-galactosyltransferase (GalT2), where RR-to-AA substitutions abolished the interaction [19]. Additionally, a single arginine residue (R11) in GnTI was sufficient for the protein localization to the Golgi. Meanwhile, another arginine residue (R2) along with a lysine residue (K5) in its cytosolic tail affected its ER-to-Golgi transportation [20].

The COPII-coated cargo carriers mediate the trafficking path of proteins, including GTs, between the ER and the Golgi. This process has been well-studied in yeast and mammalian cells; however, studies in plants are limited [28]. In Arabidopsis, AtSar1a localizes in the ER export sites (ERESs) [29] and forms a heterodimer with AtSec23a [30]. AtSar1a was involved in the trafficking from the ER to the Golgi of transcription factor bZIP28 under ER stress conditions [30], and the overexpression of the GTP-restricted mutant of AtSar1a inhibited its ER export [29]. In addition, it was demonstrated that AtSar1b localizes to Golgi-associated ERESs [29,31] and forms a protein complex with OsSec23c [31]. The RNA interference (RNAi) constructs targeting OsSar1a, OsSar1b, and OsSar1c, expressed simultaneously, suppressed the ER export of storage proteins in the rice endosperm [32]. AtSec23a, AtSec23d, AtSar1b, and AtSar1c were shown to be important for pollen development [33,34]. In the moss Physcomitrium patens (Physcomitrella), the knockout (KO) mutant of PpSec23d caused ER morphology defects and reduced ER-to-Golgi trafficking of this protein [33]. Similarly, AtSec24a was required to maintain ER morphology [35,36], and its KO mutant had a significant impact on plant growth and development [36]. All three AtSec24 paralogs are highly expressed across all Arabidopsis tissues and share similar expression patterns and subcellular localizations [37].

To test the hypothesis that specific N-terminal cytosolic motifs in Arabidopsis XXT2 and XXT5 control their ER-to-Golgi transport via COPII interactions, essential for XyGs biosynthesis in the Golgi, we examined the distinct N-terminal tails (22 residues in XXT2 and 44 residues in XXT5) to identify critical residues/motifs (e.g., di-arginine motifs) required for Golgi localization; confirmed direct protein interactions of the N-terminal cytosolic of XXTs with the COPII complex component AtSar1 or AtSec24a using bimolecular fluorescence complementation (BiFC) and pull-down assays; and validated their functional importance in plant growth and development through site-directed mutagenesis and complementation experiments in XXTs KO Arabidopsis mutant plants.

2. Results

2.1. Subcellular Localization of Truncated XXTs Transiently Expressed in Arabidopsis Protoplasts

To investigate the function of the N-terminal cytosolic tail in the XXTs’ subcellular localization and transportation from the ER to Golgi, multiple deletion mutants of XXT2 and XXT5 were generated. Proteins with N-terminal cytosolic tails truncated to different lengths were transiently co-expressed with Golgi or ER marker proteins in Arabidopsis protoplasts. Observations of the subcellular localization of truncated XXTs revealed specific protein sequences that influence their Golgi localization. As expected, XXT2 with the full length of the N-terminus was localized in the Golgi, overlapping with the Golgi marker (Figure 1). Truncated XXT2∆12A and XXT2∆15A both also localized in the Golgi, indicating that the first 15 amino acids in the N-terminus of XXT2 did not impact its localization (Figure 1 and Figure S1). Meanwhile, complete truncation of the predicted full-length N-terminus of XXT2 altered its localization. The XXT2∆N mutant displayed mainly three types of localization: ER localization, non-Golgi dot localization, and aggregation in an unidentified location in the protoplasts (Figure 1). In some protoplasts, XXT2∆N exhibited a mixture of two types of mislocalization: ER localization and non-Golgi dot localization. These results revealed that the specific sequence motif “RALRQLK” is critical for XXT2’s localization to Golgi. It is known that TMDs are essential for membrane protein localization [27,38,39,40,41,42,43]. As anticipated, partial and complete deletion of the TMD of XXT2 resulted in the aggregation of XXT2∆NM and XXT2∆M in the protoplasts (Figure 1, Figures S1A and S2).

Similar experiments using multiple deletion mutants were performed to investigate the importance of the XXT5 N-terminus for its localization. The truncation of the first 18 amino acids of the N-terminus of the XXT5 cytosolic tail did not affect its Golgi localization; XXT5∆10A and XXT5∆TTT were localized in the Golgi (Figure 2, Figures S1B and S3). Further truncated XXT5 mutants (XXT5∆di-Arg, XXT5∆40A, and XXT5∆N), which lack the specific sequence “LPTTTLTNGGGRGGR”, showed the three types of mislocalization: ER localization, non-Golgi dot localization, and aggregation (Figure 2, Figures S1B and S3). These results demonstrated that the sequence “LPTTTLTNGGGRGGR” in the N-terminal cytosolic tail of XXT5 is critical for its Golgi localization.

2.2. The Di-Arginine Motif Plays a Critical Role in Determining the Golgi Localization of XXT2 and XXT5

Earlier studies have reported that di-arginine motifs play an essential role in ER–Golgi transport of GTs and in their localization [19,22]. The revealed sequences “RALRQLK” in XXT2 and “LPTTTLTNGGGRGGR” in XXT5 contain di-arginine motifs, as emphasized in bold. To confirm the function of these di-arginine motifs in the transportation of XXTs from the ER to Golgi, the arginine residues were mutated to glutamine to maintain the side-chain structure but to eliminate the positive charge. The positive charges of arginine are believed to be critical in protein–protein interactions [44,45] and can potentially impact trafficking processes. Various mutants in the N-terminus of XXT2 were generated. The mutant with the first arginine mutated to glutamine (Q¹⁶ALR¹⁹) was named XXT2-1RQ, the mutant with the second arginine mutated (R¹⁶ALQ¹⁹) was named XXT2-2RQ, and the mutant with both arginine residues mutated (Q¹⁶ALQ¹⁹) was named XXT2-RQRQ (Figure 3A). The transient expression of XXT2-1RQ and XXT2-2RQ in the Arabidopsis protoplasts demonstrated that mutation of a single arginine in any position of the di-arginine motif resulted in mixed protein localization to both the ER and Golgi (Figure 3, Figures S4A and S5). The quantifications of XXT2-1RQ and XXT2-2RQ localization showed that about 67% were ER-localized and Golgi–ER mixed localization in the same protoplast, and another about 33% were localized in the Golgi only (Figure 3C and Figure S6A). The mutation of both arginine residues resulted in mislocalization of XXT2-RQRQ (Figure 3A), with 93.33% of the mutant protein localized either in the ER, non-Golgi dots, or a mixture of the two, whereas only 6.67% was found in the Golgi (Figure 3C, Figures S4A and S6A).

Similarly, one di-arginine motif is present in the sequence “LPTTTLTNGGGR³⁰GGR³³” of the cytosolic tail of XXT5, but another di-arginine motif (R³⁹GR⁴¹) was found closer to its TMD (Figure 4A). To investigate the function of each di-arginine motif in XXT5 localization, constructs of mutant proteins with both arginine residues mutated to glutamine in both di-arginine motifs were generated and named XXT5-RGGR and XXT5-RGR. The two arginine residues were both mutated to glutamine in the first di-arginine motifs (R³⁰GGR³³ mutated to Q³⁰GGQ³³) in XXT5-RGGR, and two arginine residues were both mutated to glutamine in the second di-arginine motifs (R³⁹GR⁴¹ mutated to Q³⁹GQ⁴¹) in XXT5-RGR (Figure 4A). These mutant proteins, XXT5-RGGR and XXT5-RGR, were localized to both the Golgi and the ER (Figure 4, Figures S4B and S7). A total of 50% of XXT5-RGGR was localized in the ER; 36.11% of XXT5-RGGR was localized both in the ER and Golgi; and 13.89% showed only Golgi localization. A total of 35.71% of XXT5-RGR was localized in the ER; 42.86% of XXT5-RGR was localized both in the ER and Golgi; and 21.43% showed the only Golgi localization. The mutation of both di-arginine motifs (four arginine residues in total) in the XXT5-RQRQ (Q³⁰GGQ³³, Q³⁹GQ⁴¹) mutant resulted in the three types of mislocalization of XXT5-RQRQ, similar to those observed in the case of mutant XXT2: 87.17% of XXT5-RQRQ was localized either in the ER, non-Golgi dots or a mixture of two, and 10.26% of XXT5-RQRQ was found in Golgi and ER-like localizations, whereas only 2.57% was found in the Golgi (Figure 4C, Figures S4B and S6B).

2.3. The Root Hair Phenotype of the Transgenic Plants Expressing Mutated XXT2 and XXT5

The delivery of XXTs to the Golgi is required for their involvement in XyG biosynthesis, and their mislocalization or ER retention significantly affect this process. To confirm this premise, the wild-type XXT2 and mutant XXT2-RQRQ were stably expressed in the xxt1xxt2 double KO mutant plants [6], while wild-type XXT5 and mutant XXT5-RQRQ were expressed in the xxt3xxt4xxt5 triple KO mutant plants [8]. Both KO double and triple mutants have somewhat similar short-root-hair phenotypes. The expression of XXT2 in the xxt1xxt2 double mutant restored the short root hairs to a length comparable with Col-0 (Figure 5A,C). The expression of XXT2-RQRQ only partially recovered the root hair length and the bubble-like extrusions at the root hairs’ tips in comparison with double-mutant plants (Figure 5A,C). Similarly, the expression of wild-type XXT5 rescued the short root hair of the xxt3xxt4xxt5 triple-mutant plants to the level of Col-0 (Figure 5B,D), while the expression of XXT5-RQRQ only partially recovered the root hair length. (Figure 5B,D).

2.4. Subcellular Localization of the Mutant XXT Proteins in the Transgenic Plants

The abnormal root hair morphology of Arabidopsis KO mutants was only partially rescued by expressing XXT2-RQRQ in xxt1xxt2 and XXT5-RQRQ in xxt3xxt4xxt5 transgenic lines. To investigate the localization of the mutant XXTs in transgenic plants, protoplasts were isolated from these plants and transfected with ER and Golgi markers for imaging (Figure 6, Figure 7 and Figures S8–S10). The images collected revealed that, while 100% of the expressed wild-type XXT2 and XXT5 proteins localized to the Golgi (Figure 6B and Figure 7B), only 11.11% of XXT2-RQRQ localized to the Golgi alone, and 12.96% showed mixed Golgi and ER-like localization (Figure 6B and Figure S6C). The remaining 75.93% were mislocalized and distributed among non-Golgi dots and ER-like locations (Figure 6B and Figure S6C). Similarly, only 27.27% of XXT5-RQRQ localized to the Golgi, including 15.15% localized to the Golgi alone (Figure 7B) and 12.12% showing mixed Golgi and ER-like localization, whereas the remaining 72.73% were distributed in ER-like and non-Golgi dot locations (Figure 7B).

2.5. The Transgenic Plants Expressing Mutant XXTs Showed a Significant Reduction in XyGs in the Cell Walls

To investigate the structural patterns of the XyG molecules in all mutant and transgenic plants, the hemicellulose fractions were prepared from their cell walls, digested with the XyG-specific hydrolase, endo-β-1,4-glucanase (XEG), and analyzed by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) [16]. As expected, XXXG, XXLG/XLXG, XXFG, and XLFG XyG subunits were observed in Col-0 (Figure 8D), whereas the xxt1xxt2 mutant plant did not contain any detectable XyG [6] (Figure 8C). The xxt3xxt4xxt5 triple mutant completely lacked the XLFG, XLLG, XXFG, and XXXG subunits, as previously reported [8] (Figure 8G). All transgenic plants expressing either wild-type or mutant XXT2 and XXT5 showed XyG subunit patterns similar to those of Col-0 plants, indicating that synthesis of complete XyG was restored to a certain extent (Figure 8A,B,E,F).

Next, the total XyG content was determined in the cell walls prepared from all the plants mentioned above. Dry cell walls were digested with Driselase, and the amount of released signature disaccharide, isoprimeverose [IP; Xyl-a-(1-6)-Glc], was quantified by high-performance anion-exchange chromatography (HPAEC). The total IP detected in all mutant and transgenic lines is expressed as a percentage relative to the amount of IP obtained from Col-0, which is set at 100% (Figure 8H). As expected, no IP was detected in the xxt1xxt2 double mutants’ cell wall [6]. In the xxt3xxt4xxt5 mutant, the XyG content was decreased to approximately 40% (Figure 8H), as previously reported [8]. The plants expressing wild-type XXT2 and XXT5 proteins showed almost complete restoration of XyG content to the level of Col-0 plants, while plants expressing mutant proteins had significantly decreased XyG content. Thus, the expression of XXT2-RQRQ in the xxt1xxt2 plants partially restored the XyG content in their cell walls from 0% to 40% compared with the KO mutant and Col-0, respectively. The expression of XXT5-RQRQ in the xxt3xxt4xxt5 plants increased XyG content by about 10% in comparison with the KO mutant (Figure 8H).

2.6. The Protein–Protein Interactions Between XXTs and Members of the COPII Complex in In Vivo (BiFC) and In Vitro Pull-Down Assays

To determine whether XXTs are cargo proteins and interact with the major COPII-coated vesicle components, a BiFC assay was used to investigate protein–protein interactions between XXTs and AtSar1/AtSec24a. BiFC signals were detected for nYFP-XXT2 with cYFP-AtSar1d and for nYFP-XXT2 with cYFP-AtSec24a (Figure 9 and Figure S11). Similarly, BiFC signals were also observed for cYFP-XXT5 with nYFP-AtSar1d and for cYFP-XXT5 with nYFP-AtSec24a (Figure 9). These results indicate that both AtSar1d and AtSec24a interact with XXT2 and XXT5.

The Sar1 and Sec23–Sec24 proteins dimer is well known to form a protein complex [46,47,48,49], and BiFC signals from their interaction with XXTs can arise simply from their spatial arrangement. Therefore, we cannot determine whether XXTs interact with AtSar1, AtSec24a, or both, because the two COPII components are in close proximity. To separate these interactions, AtSar1 and AtSec24a were individually expressed in E. coli (BL21). No prokaryotic (E. coli) COPII homologs are found in proteomic databases like UniProt (https://www.uniprot.org/) (accessed on 4 March 2026), and recombinant COPII proteins are purified using specific affinity tags to eliminate interference from contaminating host proteins, such as AtSec24a. The recombinant proteins were extracted and purified for use as bait for in vitro pull-down assays with full-length membrane proteins XXT2 and XXT5 obtained as total membrane protein extract from Arabidopsis plants. The primary anti-Myc antibody was used to detect AtSar1, the anti-HA antibody to detect AtSec24a, and the anti-GFP antibody to detect CFP-tagged XXT2 or XXT5. These experiments showed that full-length XXT2 and XXT5 interacted with AtSar1b, AtSar1c, and AtSar1d, whereas no interaction with AtSec24a was detected (Figure 10).

The cytosolic tails of XXTs are exposed to the cytosol and their stem regions and catalytic domains are located in the Golgi lumen. This positioning makes the cytosolic tails the most likely region to interact with soluble AtSar1. The peptides of the N-terminal cytosolic tails of XXT2 and XXT5 were synthesized by PEPTIDE 2.0 Inc. (Chantilly, VA, USA). A Flag tag was fused to each peptide, which was composed of the cytosolic tails of wild-type XXT2 and XXT5, with additional peptides with R-to-Q mutations (named XXT2Q) or R-to-A (named XXT2A) mutations. Two additional mutant peptides (R-to-A) were used to increase the mutation’s potential impact. First, it was confirmed that the amounts of wild-type peptides of XXTs and peptides with mutations used in the following pull-down experiments were comparable (Figure 11B–G, shown in the line corresponding to each peptide). Since the AtSar1 proteins lacked a Flag tag, they did not show any nonspecific interaction with anti-Flag resin and therefore served as negative controls (Figure 11B–G). The COPII coat protein AtSar1b showed protein–protein interactions with the cytosolic tails of wild-type XXT2 and XXT5 (Figure 11B,C). The mutation from R to Q in the di-arginine motifs significantly affected the protein–protein interactions between the cytosolic tails of XXT5 and AtSar1b (Figure 11C) but did not affect the interaction between XXT2 and AtSar1b (Figure 11B). However, the peptides with R-to-A mutations showed significantly weaker interactions with AtSar1b for both XXT2 and XXT5 (Figure 11B,C). The peptides of both XXTs also interacted with AtSar1c and AtSar1d (Figure 11D–G). However, the strength of these interactions was somewhat weaker than that with AtSar1b. Both types of mutations, either R-to-Q or R-to-A, significantly decreased the protein–protein interactions between both XXTs and AtSar1c or AtSar1d (Figure 11D–G). The experiments further confirmed the interactions between XXT peptides and the AtSar1 protein, with no interactions observed with AtSec24a (Figure S12).

3. Discussion

3.1. Role of Cytosolic Tails and TMDs

Correct subcellular localization of proteins is essential for their proper function within cells. Golgi-localized GTs must be delivered and retained in the Golgi after being properly folded in the ER. The importance of N-terminal cytosolic tails and TMDs of GTs for their transport and localization has been reported earlier [19,20,21,22,24,28,40,50]. In this study, we investigated the contribution of cytosolic tails and TMDs to Golgi localization using two XXTs involved in plant polysaccharide biosynthesis. The deletion of N-terminal cytosolic tails of XXT2 and XXT5 altered their Golgi-localization (Figure 1, Figure 2 and Figure S1), indicating the possible critical signaling role of their N-termini in ER-to-Golgi trafficking. Furthermore, deletion of both the N-terminus and TMD of XXT2 caused protein aggregation (Figure 1 and Figure S1), indicating the indispensable role of TMDs in proper GT trafficking via the secretory pathway.

3.2. Role of Di-Arginine Motifs

By gradually deleting the cytosolic tails of XXTs, we identified the specific protein sequences containing di-arginine motifs critical to their subcellular localization. Subsequent confirmation through the mutation of arginine residues showed that di-arginine motifs are primarily responsible for the proper Golgi localization of XXTs. Previous studies have reported the function of arginine residues and di-arginine motifs in other proteins’ localization and trafficking [19,20,24,27,51,52]. In our study, R-to-Q mutations of the di-arginine motifs resulted in three distinct patterns of mislocalization in 90% of the mutated XXTs: ER-like, non-Golgi dot localization, and aggregation (Figure 3, Figure 4, Figures S4 and S6). This mislocalization of XXTs impeded the enzymes’ function in XyG biosynthesis. Total XyG content within the cell wall of transgenic plants expressing mutated XXTs was significantly reduced compared with Col-0 and KO mutants expressing corresponding wild-type XXTs (Figure 8). Moreover, expression of mutated XXTs resulted in significantly shorter root hairs compared to Col-0 plants (Figure 5). These results stipulate the critical role of the di-arginine motif in XXTs protein trafficking and function.

Notably, the di-arginine motifs in XXT2 and XXT5 exhibited a cooperative effect on protein localization. In the XXT2-1RQ and XXT2-2RQ mutants, substitution of a single arginine residue within the motif triggered mislocalization, with about 33% of the mutated XXT2 localized in the Golgi only, and about 23% of the mutated XXT2 localized both in the ER and Golgi (Figure 3 and Figure S6). The mutation of both arginine residues caused greater mislocalization, with only 7% of XXT2-RQRQ remaining in the Golgi (Figure 3 and Figure S6). Two di-arginine motifs were found in the cytosolic tail of XXT5, and both exerted influence on the subcellular localization of XXT5. The mutation of all four arginine residues of both di-arginine motifs significantly increased the mislocalization of XXT5-RQRQ (Figure 4 and Figure S6). Our observations are consistent with previous findings that the mutation of any single arginine residue in the di-arginine motifs of Golgi-localized GalNAcT causes a mixed distribution between the Golgi and ER [19], whereas the mutation of both arginine residues results in ER localization.

3.3. Non-Golgi Dot Structures

In our study, we also observed the localization of mutant XXTs in non-Golgi dot structures. This type of localization was identified by the lack of overlap between the fluorescent signals of the Golgi marker and mutant XXTs; however, some XXT signals showed visible adherence to the ER marker and were surrounded by the ER (Figure 1, Figure 2, Figure 3 and Figure 4). We speculate that the mislocalized mutant XXTs were retained in the ER, specifically at ERESs where COPII-coated vesicles are assembled [53,54,55,56]. Cargo protein sorting, a critical step in protein trafficking, occurs during COPII vesicles’ assembly. The plant Sec16 and MAG5 proteins have been reported to be localized at ERESs in a previous study [57,58]. In that study, MAG5 was surrounded by a cup-shaped subdomain of the ER [58]. In our study, the non-Golgi dot localization of the mutant XXT5-RQRQ was similar to that observed for the MAG5 protein (Figure 4B). In the future, it would be interesting to co-express MAG5 and mutant XXTs together to determine the precise locations of the observed non-Golgi dots.

3.4. Residual Golgi Trafficking Persists

Although mutation of the di-arginine motifs altered the localization of most XXTs, a small percentage of the XXT2-RQRQ and XXT5-RQRQ mutant proteins was still found in the Golgi; less than 10% of XXT2-RQRQ and 5% of XXT5-RQRQ were still delivered to the Golgi when transiently expressed in Arabidopsis protoplasts (Figure 3, Figure 4, Figures S4 and S6). A slightly higher fraction of mutant proteins was observed in the Golgi when constitutively expressed in KO plants: about 11% of XXT2-RQRQ and about 15% of XXT5-RQRQ (Figure 6, Figure 7, Figures S6 and S8). MALDI-TOF analysis demonstrated that mature, fully branched XyGs structures were synthesized in all transgenic plants, including XXTs with mutated di-arginine motifs (Figure 8A,B,E,F). This could indicate that a small fraction of XXTs delivered to the Golgi are sufficient to support some level of XyG synthesis. Such partial restoration of XyG in mutant cell walls was also sufficient for the partial recovery of root hair phenotypes normally observed in KO plants (Figure 5A,B). Since the substitution of both arginine residues does not completely inhibit the delivery of XXTs to the Golgi, it is plausible that additional residues or motifs likely contribute to their ER-to-Golgi trafficking. For example, it has previously been shown that N-terminal cooperation with the TMDs/catalytic domains of GTs can contribute to their localization and transport [21,22,39]. Mutation of the di-arginine motif in the cytosolic tails of XXTs likely slowed their delivery rate but did not eliminate their presence in the Golgi. For example, in yeast, it demonstrated that impairing the ER export signal significantly decreases the transport rate and impedes the export of cargo proteins to the Golgi [19,25].

3.5. Direct XXT-AtSar1 Interactions Confirmed

In yeast, mammals, and plants, COPII-coated vesicles transport cargo from the ER to the Golgi. Specific motifs in cargo proteins have been reported to interact with components of the COPII coat complex, including Sec24 and Sar1. In yeast and mammalian cells, Sec24 functions in cargo selection through direct protein–protein interactions with cargo proteins or cargo receptors [59,60,61,62,63]. Though the BiFC assay provided signals suggesting that XXTs interact with both AtSar1d and AtSec24a (Figure 9), this technique has limitations. As Sar1 interacts with the Sec23–Sec24 heterodimer to form a complex [46,47,48,49], their close spatial proximity likely means that XXTs bound to either partner may bring the split fluorophore fragments together, producing fluorescence even when one of the apparent interactions is indirect. This spatial coupling can therefore generate false-positive signals that do not necessarily reflect direct binding between XXTs and either COPII component. Therefore, the BiFC results should be corroborated by independent approaches, for example, by pull-down experiments.

To avoid the impact of Sar1 and Sec23–Sec24 protein complex formation on the interpretation of the results, AtSar1b, AtSar1c, AtSar1d and AtSec24a were expressed separately in E. coli BL21 cells as Myc- or HA-tagged recombinant proteins. Total membrane protein extracts were prepared from young Arabidopsis seedlings expressing GFP-tagged XXT2 or XXT5 proteins. These pull-down experiments showed that XXTs interact with AtSar1b, AtSar1c and AtSar1d but not with AtSec24a (Figure 10). These results strongly suggest that XXTs interact directly with AtSar1 proteins rather than with AtSec24a. To further understand which region of XXTs mediates these interactions, synthetic peptides of XXT cytosolic tails were used in pull-down assays, which confirmed that the cytosolic tails interact with AtSar1 and not with AtSec24a (Figure 11 and Figure S12).

3.6. Motif Mutations Weaken Direct XXT-AtSar1 Interactions

In Arabidopsis, there are five paralogs of AtSar1 proteins, and it was proposed that their functions do not completely overlap [26,29,30,33,64,65,66,67]. Here, we studied the roles of AtSar1b, AtSar1c, and AtSar1d in the potential interaction with XXTs. These three Sar1 paralogs were selected due to the tissue-specific expression overlaps with that of the XXTs, whereas AtSar1a is highly expressed only in siliques, and AtSar1e is primarily expressed during seed germination (https://www.arabidopsis.org/) (accessed on 4 March 2026). Results from BiFC assays, peptide-based pull-down assays, and full-length protein pull-down assays consistently demonstrated that XXTs interact with AtSar1b, AtSar1c, and AtSar1d (Figure 9, Figure 10 and Figure 11).

However, the XXTs exhibited different levels of interactions with AtSar1b, AtSar1c and AtSar1d, showing stronger interactions with AtSar1b compared to AtSar1c and AtSar1d (Figure 11). The substitution of R to Q residues in the di-arginine motif of XXTs weakened their interactions with AtSar1 paralogs, and the substitution of R to A impaired the interaction even further (Figure 11). The observed differences among AtSar1b, AtSar1c, and AtSar1d in their ability to interact with XXTs may imply that different AtSar1 paralogs might have diverse contributions to trafficking GTs in plants. Di-arginine motifs are found in the cytosolic tails of XXT1, XXT2, XXT5, XLT2, MUR3, and FUT1, the other GTs involved in XyG biosynthesis. Based on previous studies on the function of the AtSar1 protein in cargo sorting [19,20,25,27,67] and our data from this study, we hypothesize that the AtSar1 proteins are the primary contributors in the ER-to-Golgi trafficking of GTs involved in polysaccharide synthesis. Our results demonstrate that plant Golgi-resident membrane proteins such as GTs interact with the COPII complex via AtSar1 proteins, and arginine residues in their N-terminal cytosolic tails are critical for this interaction, which determines the effectiveness of their delivery from the ER to the Golgi.

3.7. Future Perspectives

There are numerous GTs localized in the Golgi, and not all have a di-arginine motif in their cytosolic N-termini. Therefore, it will be very informative to investigate other GTs to reveal their specific motifs involved in interaction with members of the COPII complex. Also, it is unclear whether all GTs are recruited into COPII-coated vesicles via Sar1 or Sec24. More detailed investigation of the structural aspects of interactions between GTs and member proteins of the COPII complex would increase our understanding of the mechanisms of GTs trafficking and their control, which, in turn, will broaden our understanding of polysaccharide biosynthesis and what controls its reducibility.

4. Conclusions

In this study, we have provided substantial evidence suggesting that the di-arginine motifs in the N-termini of two XXTs proteins are critical for their efficient transport from the ER to the Golgi and most likely are involved in interactions with COPII complex member Sar1. Our results also strongly suggest that Sar1, not Sec24, is the main recruiter of XXTs into COPII coated vesicles. The results of this study increase our understanding of the mechanisms underlying the delivery of polysaccharide-synthesizing GTs from the ER to the Golgi and confirm that their proper localization is required for their functioning.

5. Materials and Methods

5.1. Plant Growth and Plasmid Construction

The xxt1xxt2 double mutant [6] and xxt3xxt4xxt5 triple mutant [8] were obtained earlier. To observe the root hair phenotype, sterilized seeds were grown on Petri dishes in ½ MS media (pH 5.8) with 1% (w/v) agar under long-day conditions (16 h:8 h, light:dark) at 22 °C. Seeds were germinated in horizontally positioned dishes, which were turned vertically later to ensure the roots grew straight.

To generate multiple mutant XXTs, gene-specific primers (Supplemental Table S1) were utilized to clone the two fragments of truncated XXTs from previous vectors [17,23], and the fusion PCR was applied to link the two fragments as the full length of truncated XXTs. Phire Hot Start II DNA Polymerase (Thermo Fisher Scientific Inc., Waltham, MA, USA, F122) was used in the fusion PCR, which used the following: 5× Phire Reaction Buffer 10 μL; forward primer (2 µM) 12.5 μL; reverse primer (2 µM) 12.5 μL; 10 mM dNTPs 1 μL; DMSO 1.5 μL; Fragment 1 (20 ng/μL) 2 μL; Fragment 2 (20 ng/μL) 2 μL; Phire Hot Start II DNA Polymerase 1 μL; and double-distilled water (ddH₂O) added to 50 μL. The cycling steps were 98 °C for 5 min, followed by 35 cycles (98 °C for 10 s, 70 °C for 30 s, 67 °C for 30 s, 64 °C for 30 s, and 72 °C for 1 min), then 72 °C for 5 min, and finally 12 °C for ∞. The fusion PCR products were digested with the corresponding restriction enzymes and inserted into pUBN-CFP expression vectors. For the BiFC assay and confocal microscopy study, all genes (AtXXT2, AtXXT5, AtSar1d, AtSec24a and α-Mannosidase (MNS1)) were cloned with gene-specific primers from previous vectors [17,23] and full-length complementary DNA (cDNA). The PCR products were digested with the corresponding restriction enzymes and inserted into either the N- or C-terminal fragment of YFP (nYFP and cYFP) at the N-terminal [17,23] in pSAT vectors [68].

5.2. Transgenic Arabidopsis Plants

Recombinant Agrobacterium with pUBN::CFP-XXT2 (XXT2) or pUBN::CFP-XXT2-RQRQ (XXT2-RQRQ) was transformed into the xxt1xxt2 double mutant, and Agrobacterium with pUBN::CFP-XXT5 (XXT5) or pUBN::CFP-XXT5-RQRQ (XXT5-RQRQ) was transformed into the xxt3xxt4xxt5 triple mutant by the floral dipping method [69]. Sterilized transformants were selected on the Petri dishes with 250 µg/mL Hygromycin B (Thermo Fisher Scientific Inc., Waltham, MA, USA, 10687010), and the expression of CFP in the transgenic seedlings was confirmed by detecting the CFP fluorescence under the fluorescent microscope. The protein expression levels of CFP–XXT2, CFP–XXT2-RQRQ, CFP–XXT5, and CFP–XXT5-RQRQ were examined by Western blot analysis (Figure S13).

5.3. Transient Expression in Arabidopsis Protoplasts

The recombinant pUBN::CFP-XXT, pUBN::CFP-truncated XXT (including XXT2Δ12A; XXT2Δ15A; XXT2ΔN; XXT2ΔNM; XXT2ΔM; XXT5Δ10A; XXT5ΔTTT; XXT5Δdi-Arg; XXT5Δ40A and XXT5ΔN) and pUBN::CFP-mutated XXT (including XXT2-1RQ; XXT2-2RQ; XXT2-RQRQ; XXT5-RGGR; XXT5-RGR; and XXT5-RQRQ) vectors were extracted from E. coli cells and were purified via PureLink™ HiPure Plasmid Maxiprep Kit (Invitrogen Inc., Carlsbad, CA, USA, K210007). The preparation of Arabidopsis protoplasts from Col-0 and transgenic Arabidopsis, as well as the transient expression of constructs, was performed according to a previously described method [8,17,23]. We took about 20–30 leaves from 4 to 6 weeks-old Arabidopsis plants. Then, we placed the adhesive tape with the sticky side facing up and then attached the leaves to the sticky side. After all the leaves were attached, we used another piece of tape to stick to the unattached side of the leaves, so that the leaves were tightly sandwiched between the two pieces of tape. Then, we carefully peeled off one side of the tape; several layers of cells on the surface of the leaf were removed by the tape, exposing the inner layers of cells. We cut the piece of tape containing most of the leaf into small sections and placed them in a tube, ensuring that the enzyme solution (0.25% [w/v] Macerozyme, 1.0% Cellulase, 0.4 mm mannitol, 8 mm CaCl₂, 5 mm MES-KOH, pH 5.6, and 0.1% bovine serum albumin) completely immerses the tape with the leaf fragments. Incubated leaf fragments in 20 mL of enzyme solution for 1.5 h in the dark with gentle agitation at 70 rpm. After incubation, suspended protoplasts were overlaid on 20 mL of 21% (w/v) sucrose solution, and centrifuged at 400× g for 5 min. The collected protoplasts were washed with 10 mL of W5 solution (154 mM NaCl, 125 mM CaCl₂, 5 mM KCl, 5 mM glucose, and 1.5 mM MES-KOH, at pH 5.6) three times and pelleted again by centrifugation at 400× g for 5 min. Protoplasts were then resuspended in 2 mL of W5 solution, and their concentration was determined using a hemocytometer (0.1 mm depth). After overnight incubation in the dark at room temperature, the fluorescence signals of CFP, YFP, and mCherry were detected using a confocal microscope (Leica SP5 system and Leica STED super-resolution, Leica Microsystems Inc., Deerfield, IL, USA) [8].

For the BiFC assay, pSAT vectors with different genes were purified and transformed into Arabidopsis protoplasts. After two days of incubation in the dark at room temperature, fluorescence signals were checked with the same confocal setting [17,23].

5.4. Quantification of Localization of XXTs and Mutant XXTs in Arabidopsis Protoplasts

To quantify the subcellular localization of XXTs and mutant XXTs relative to the Golgi marker, more than 100 protoplasts were randomly selected per protein. The localization patterns of XXTs and mutant XXTs were classified into five categories: (1) Golgi, where more than 50% of the XXT or mutant XXT signals overlapped with the Golgi marker and no ER-like signal was observed; (2) Golgi + ER-like, where more than 50% of the signal overlapped with the Golgi marker and additional ER-like signal was present; (3) ER-like, with an ER-like signal only; (4) ER-like + non-Golgi dots, with an ER-like signal and more than 50% of the remaining XXT or mutant XXT signals appearing as dot-like structures that did not overlap with the Golgi marker; and (5) non-Golgi dots, where more than 50% of the XXT or mutant XXT signals appeared as dot-like structures that did not overlap with the Golgi marker and no ER-like signal was observed. The experiments were independently repeated three times. The percentage of protoplasts in each category and the corresponding standard deviation (SD) were then calculated.

5.5. Cell Wall Extraction and Analysis of Hemicellulose Fractions

Alcohol-insoluble cell wall material and the hemicellulose fractions from all studied plant lines and Col-0 were extracted using established standard protocols [8,16]. The three solutions and buffers (50 mM CDTA: 50 mM ammonium oxalate (1:1) buffer, 50 mM Na₂CO₃, and 4 N KOH) were sequentially applied in the three different polysaccharide fractions extraction. The hemicellulose fractions digested with XEG were analyzed by MALDI-TOF mass spectrometry. The protocol followed was published in [8]. The Na₂CO₃ + 4 N KOH hemicellulose fraction was resuspended in 25 mM sodium acetate with purified recombinant XEG (EC 3.2.1.151) and incubated at 37 °C for 17 h. Dihydroxybenzoic acid matrix solution and standard or digested sample were spotted onto a 96-well MALDI-TOF plate (Kratos Analytical, Manchester, UK) and allowed to dry. For each sample, 100 profiles were collected in reflection mode at a laser power setting of 140.

The same hemicellulose fractions were separately digested with Driselase, a mixture of exoglycosidases and endoglycosidases (Sigma-Aldrich, St Louis, MO, USA, D9515), and the total amount of a released signature disaccharide, IP (Xyl-a-(1-6)-Glc), was quantified by HPAEC using the gradient conditions described previously [8,16]. Lyophilized hemicellulose fraction was resuspended in 25 mM sodium acetate containing 0.2 µg mL⁻¹ Driselase and incubated at 37 °C overnight with shaking. Supernatants were collected by centrifugation and analyzed by HPAEC using previously described gradient conditions.

5.6. Protein Extraction

About 15 g of 14-day-old young Arabidopsis seedlings expressing CFP-XXT2 and CFP-XXT5 were collected and ground in liquid nitrogen. A total of 60 mL protein extraction buffer (50 mM HEPES, 0.3 M sucrose, 65 mM NaCl, and 3 mM EDTA, at pH 7.5) supplemented with 3 mL protease inhibitor cocktail (1 mM E-64, 1 mM leupeptin, 100 mM AEBSF, 100 mM benzamidine, and 100 mM PMSF in methanol) was added. Plant debris was removed by filtration through miracloth (Millipore Sigma, Lenexa, KS, USA, 4758551R), and the clarified filtration was transferred to new 50 mL centrifuge tubes. The filtration was centrifuged at 20,000× g for 1 h at 4 °C, and the supernatant was transferred to ultracentrifuge tubes (Beckman Coulter Life Sciences, Indianapolis, IN, USA, #355618). The supernatant was then ultracentrifuged at 100,000× g for 1 h at 4 °C using a Ti-70 rotor. Then, the supernatant was discarded, and the pellet was washed with 1 mL protein extraction buffer. After the three washes, the membrane pellet was resuspended in 500 µL of protein extraction buffer containing 2% Triton X-100 and 1% NP-40 and incubated with shaking at 4 °C overnight to solubilize the membrane-bound proteins. The solubilized membrane fraction was transferred to 1.5 mL tubes, kept on ice, and sonicated to further enhance solubilization. Finally, the samples were centrifuged at 14,000× g 10 min and 4 °C to remove the insoluble remaining aggregation, and the supernatants (containing the total membrane protein) were collected for subsequent analysis.

5.7. Protein Purification

The cDNA of the COPII complex proteins AtSar1b (AT1G56330), AtSar1c (AT4G02080), AtSar1d (AT3G62560) and AtSec24a (AT3G07100) were inserted into the pET His6 Sumo TEV LIC cloning vector containing an N-terminal 6xHis tag using gene-specific forward and reverse primers and expressed in E. coli (BL21). Transformed E. coli was incubated in 500 mL of lysogeny broth medium at 37 °C, and when cells reached a density with OD of 0.4–0.6, the temperature was reduced to 18 °C, isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM, and the cells were incubated for another 18 h at 18 °C. E. coli cells were collected by centrifuge at 6000× g for 10 min. Collected cells were resuspended in the cold lysis buffer (25 mM Tris; pH 7.4, 300 mM NaCl, 0.5 mM EDTA) and treated by five freeze/thaw cycles in liquid nitrogen. After thawing, the lysozyme solution was added to achieve a final concentration of 1 mg/mL, and the cells were incubated at 4 °C for 30 min to 60 min. The cells were sonicated five times for 15 s each and centrifuged at 20,000× g for 30–60 min to collect soluble proteins. The soluble protein fractions were passed through HisPur Ni-NTA Resin (Thermo Fisher Scientific Inc., Waltham, MA, USA, 88223) to purify the recombinant proteins, following the manufacturer’s protocol. Amicon Ultra-15 Centrifugal Filter units (Millipore Sigma, Lenexa, KS, USA, UFC901024 and UFC903024) were used to purify the proteins further. Meanwhile, we gradually added about 15 mL to 20 mL of lysis buffer (25 mM Tris; pH 7.4, 300 mM NaCl, 0.5 mM EDTA) to replace the elution buffer (50 mM Tris-HCl, 150 mM NaCl, and 300 mM imidazole). The purified COPII complex proteins AtSar1b, AtSar1c, AtSar1d, and AtSec24a were checked by the Coomassie stains and Western blot (Figure S14). Purified proteins were stored in the lysis buffer and were used for the pull-down assay in vitro.

5.8. In Vitro Pull-Down Assay

For the in vitro pull-down assays with full-length XXT2 and XXT5, we used 400 μg of AtSar1 or AtSec24a purified proteins, incubated with 100 μL of HisPur™ Ni-NTA Resin (Thermo Fisher Scientific Inc., Waltham, MA, USA, 88223) in each reaction at 4 °C for 1 h with end-to-end shaking. After incubation, the resin in the column was washed five times with Wash Buffer (pH 7.4, 20 mM sodium phosphate, 300 mM sodium chloride, and 25 mM imidazole). After washing, the resin was incubated with 500 µL total membrane protein solution at 4 °C for 4 h with end-to-end shaking. The same washing steps were repeated as described above, and proteins were eluted with 160 μL acid elution buffer (pH 7.4, 20 mM sodium phosphate, 300 mM sodium chloride, and 250 mM imidazole). The collected elution fractions were analyzed by Western blotting using tag-specific antibodies. The full lengths of XXT2 and XXT5 proteins were individually incubated with HisPur™ Ni-NTA Resin (Thermo Fisher Scientific Inc., Waltham, MA, USA, 88223) as the negative control, and the results are shown in Figure S15.

The cytosolic N-end tails of XXT2 and XXT5 fused with Flag tag were synthesized by PEPTIDE 2.0 Inc. (Chantilly, VA, USA). The sequences of the synthesized peptides are as follows:

• XXT2: MIERCLGAYRCRRIQRALRQLKDYKDDDDK.
• XXT2Q: MIERCLGAYRCRRIQQALQQLKDYKDDDDK.
• XXT2A: MIERCLGAYRCRRIQAALAQLKDYKDDDDK.
• XXT5: MGQDGSPAHKRPSGSGGGLPTTTLTNGGGRGGRGGLLPRGRQMQKTFNNIKDYKDDDDK.
• XXT5Q: MGQDGSPAHKRPSGSGGGLPTTTLTNGGGQGGQGGLLPQGQQMQKTFNNIKDYKDDDDK.
• XXT5A: MGQDGSPAHKRPSGSGGGLPTTTLTNGGGAGGAGGLLPAGAQMQKTFNNIKDYKDDDDK.

In in vitro pull-down assays with cytosolic N-end tails of XXT2 and XXT5, Pierce Anti-DYKDDDDK Affinity Resin (Thermo Fisher Scientific Inc., Waltham, MA, USA, PIA36801) and the company protocol were used. For each reaction, 50 μL of resin was washed three times with cold lysis buffer (25 mM Tris, pH 7.4, 300 mM NaCl, and 0.5 mM EDTA) and incubated with 40 μg of peptides with a Flag tag dissolved in 300 μL of lysis buffer at 4 °C for 4 h with end-to-end shaking. After incubation, the resin in the column was washed twice with Phosphate-Buffered Saline buffer and once with ddH₂O. After washing, the resin was incubated with 4 nmol of each COPII complex protein and dissolved in lysis buffer at 4 °C overnight with end-to-end shaking. The same washing steps were repeated as described above, and proteins were eluted with 200 μL of acid elution buffer (0.1 M glycine; pH 2.8). The acid was then immediately neutralized by adding neutralization buffer (1 M Tris; pH 8.5). Collected elution fractions were analyzed by Western blotting using tag-specific antibodies. AtSar1 and AtSec24a proteins were individually incubated with Pierce Anti-DYKDDDDK Affinity Resin as the negative control, and the results are shown in Figure S16.

5.9. Western Blotting

The SDS-PAGE with an 18% acrylamide gel was used for the separation of the peptides with a Flag tag, while a 10% acrylamide gel was used for the separation of AtSar1, and 8% acrylamide gel was used for the separation of XXTs and AtSec24A recombinant proteins. After SDS-PAGE separation, the proteins were transferred to nitrocellulose membranes (0.2 mm; Bio-Rad, Hercules, CA, USA) for immunodetection. DYKDDDDK Tag monoclonal antibodies (Invitrogen Inc., Carlsbad, CA, USA, MA191878) were used (1:1000 dilution) to detect synthesized XXT peptides. Anti-Myc Tag antibodies (Sino Biological, Chesterbrook, PA, USA, 50-161-0309) were used (1:2000 dilution) to detect the AtSar1 proteins. HA Tag monoclonal antibodies (Invitrogen Inc., Carlsbad, CA, USA, PI26183) were used (1:2000 dilution) to detect the AtSec24A proteins. Monoclonal anti-GFP antibodies (Covance, Princeton, NJ, USA) MMS-118P were used (1:6000 dilution) to detect the CFP-XXT fusion proteins. The membranes were incubated with West Pico PLUS Chemiluminescent Substrate (Super Signal, Englewood, FL, USA) and visualized using a ChemiDocXRS+ (Bio-Rad) system (Bio-Rad, Hercules, CA, USA). The intensities of the bands on the membrane were quantified using ImageJ.