MAG2 and MAL Regulate Vesicle Trafficking and Auxin Homeostasis With Functional Redundancy

Auxin is a central phytohormone and controls almost all aspects of plant development and stress response. Auxin homeostasis is coordinately regulated by biosynthesis, catabolism, transport, conjugation, and deposition. Endoplasmic reticulum (ER)-localized MAIGO2 (MAG2) complex mediates tethering of arriving vesicles to the ER membrane, and it is crucial for ER export trafficking. Despite important regulatory roles of MAG2 in vesicle trafficking, the mag2 mutant had mild developmental abnormalities. MAG2 has one homolog protein, MAG2-Like (MAL), and the mal-1 mutant also had slight developmental phenotypes. In order to investigate MAG2 and MAL regulatory function in plant development, we generated the mag2-1 mal-1 double mutant. As expected, the double mutant exhibited serious developmental defects and more alteration in stress response compared with single mutants and wild type. Proteomic analysis revealed that signaling, metabolism, and stress response in mag2-1 mal-1 were affected, especially membrane trafficking and auxin biosynthesis, signaling, and transport. Biochemical and cell biological analysis indicated that the mag2-1 mal-1 double mutant had more serious defects in vesicle transport than the mag2-1 and mal-1 single mutants. The auxin distribution and abundance of auxin transporters were altered significantly in the mag2-1 and mal-1 single mutants and mag2-1 mal-1 double mutant. Our findings suggest that MAG2 and MAL regulate plant development and auxin homeostasis by controlling membrane trafficking, with functional redundancy.


INTRODUCTION
Auxin is a central phytohormone for almost all aspects of plant growth and development (reviewed in Gomes and Scortecci, 2021), and response to environmental stimuli (reviewed by Zhao, 2018). Auxin homeostasis regulated by coordination of auxin biosynthesis, catabolism, transport, conjugation, and deposition optimizes plant development and adaption to environmental stress (Bhalerao and Bennett, 2003;Blakeslee et al., 2019). Auxin gradients determine developmental outcomes (Leyser, 2005;Habets and Offringa, 2014;Zhao, 2018). Both roots and shoots exhibit auxin gradients across longitudinal axes, and auxin levels are generally most concentrated in organ meristems and rapidly dividing tissues (Kramer and Bennett, 2006). Auxin transport is controlled mainly by AUXIN1 (AUX1), PIN-FORMED (PIN), and PIN-LIKES (PILS) family carriers. These proteins coordinately control auxin intercellular and intracellular transport and determine plant morphogenesis (Mravec et al., 2009;Barbez et al., 2012). Canonical PIN proteins such as AtPIN1-4 and AtPIN7 localize in the plasma membrane (PM) asymmetrically and play an overarching role in plant development by regulating directional cell-to-cell auxin transport (reviewed by Naramoto, 2017;Béziat and Kleine-Vehn, 2018). PILS proteins are observed to localize only in the endoplasmic reticulum (ER) (Barbez et al., 2012;Sauer and Kleine-Vehn, 2019), while, non-canonical PINs display diverse localization. For instance, AtPIN5 exhibits cell type-dependent localization, at the PM in aerial tissues and intracellular localization in root vascular cells (Ganguly et al., 2014); AtPIN6 shows dual localization in the ER and the PM (Simon et al., 2016;Ditengou et al., 2018); PIN8 is colocalized with PIN5 in the ER in pollen (Ding et al., 2012). Noncanonical PIN and PILS proteins likely sequester auxin in the ER and have an impact on cellular auxin signaling and homeostasis (Mravec et al., 2009;Barbez et al., 2012;Béziat et al., 2017;Middleton et al., 2018;Feraru et al., 2019;Sun et al., 2020).
After being synthesized and assembled in the ER (Borgese et al., 2006), canonical PIN proteins are delivered to the PM through the secretory pathway, and they maintain their homeostasis in the PM by the cycling machinery (Naramoto, 2017). Phosphorylation of PIN proteins, which appears to control both PIN directional delivery and activities, is regulated by kinases, D6 protein kinases (D6PKs), PINOID (PID), wavy root growth (WAG)1, WAG2, and protein phosphatese 2A (PP2As) (Friml et al., 2004;Michniewicz et al., 2007;Dhonukshe et al., 2010;Zourelidou et al., 2014;Weller et al., 2017;Barbosa et al., 2018;Zhou and Luo, 2018). The impact of PID and PP2As on PIN phosphorylation status determines PIN cycling and maintains PIN polar localization (Máthé et al., 2021). The vesicle trafficking system maintains organelle identities and homeostasis to contribute to proper cellular activities. Recognition machineries of a donor with target membranes consist of tethering factors, Ras-related in brain (RABs), ADPribosylation factors (ARFs), guanine nucleotide exchange factors (GEFs), etc. (Lamber et al., 2019;Homma et al., 2021). Tethering factors mediate the first contact between arriving vesicles and target membrane (Grosshans et al., 2006), and transfer the machinery to downstream factors such as soluble N-ethylmalemide sensitive factor attachment protein receptors (SNAREs) (Wang et al., 2017). SNAREs facilitate membrane fusion of transport vesicles with target membranes. According to sequences of center amino acids in the SNARE motif, SNARE proteins are classified into Q-SANREs (including Qa-, Qb-, and Qc-SNAREs) and R-SNAREs. Specific combination of R-with Q-SNAREs forms a SNARE complex to drive membrane fusion (Fasshauer et al., 1998).
Despite the important regulatory roles of MAG2 in vesicle transport and stress response, mag2 mutants just exhibit mild developmental abnormalities. It is reported that MAG2 has a homolog protein, MAG2-like (MAG2-Like (MAL), At1g08400) (Zhao et al., 2013;Zhao and Lu, 2014). We isolated a T-DNA insertion mutant, mal-1, and found that it also had slight developmental phenotypes. In order to analyze MAG2 and MAL function in plant development, we generated a double mutant, mag2-1 mal-1. As expected, the mag2-1 mal-1 double mutant had serious developmental defects such as decreased germination activities, dwarf and partial seed abortion, and abnormal response to salt and osmotic and ABA stress. SSP precursors also accumulated at a higher level in the double mutant seeds than in the single mutant seeds. Proteomic analysis revealed that signaling, metabolism, and stress response were affected in mag2-1 mal-1, especially membrane trafficking, auxin biosynthesis, signaling, and transport. Biochemical and cell biological analysis indicated that the mag2-1 mal-1 double mutant had more serious defects on vesicle transport than the single mutants. Auxin distribution and auxin transporter accumulation were significantly altered in mag2-1, mal-1, and mag2-1 mal-1. Our findings suggested that MAG2 and MAL regulate plant development, auxin homeostasis, and stress response potentially by controlling vesicle trafficking, and that they are functionally redundant.

Plant Materials and Growth Conditions
Arabidopsis thaliana ecotype Col-0 was used as wild-type plants. The T-DNA-tagged line (mal-1, GABI_kat_288E12) was provided by the Arabidopsis Biological Resource Center (ABRC) at Ohio State University. The mag2-1 mutant was from our previous study (Li et al., 2006(Li et al., , 2013Zhao et al., 2018). Homozygous plants were obtained by PCR screening using genespecific primers. Arabidopsis seeds were surface-sterilized and sown either on soil or in 0.8 or 1.2% agar with 1/2 Murashige and Skoog medium (PhytoTech, China) and 1% (w/v) sucrose. The plants were grown at 22 • C under 16: 8-h/light: dark cycles.

RNA Extraction and RT-PCR Analysis
Total RNA was isolated using RNAiso Plus (9109; TAKARA, Japan). Total RNA 0.5-1 µg was reverse transcribed using PrimeScript TM RT Master Mix (Perfect Real Time) (RR036A; TAKARA, Japan). Semiquantitative RT-PCR and RT-qPCR were performed according to the manufacturer's instructions. ACT2 was used as an endogenous control for RT-PCT.

Yeast Two-Hybrid Assay
For the yeast two-hybrid assay, AtSYP81, AtSEC20, and MAG2 constructs were generated as described in our previous study (Li et al., 2006). The cDNA of MAL was amplified and fused in-frame downstream of the GAL4 activation domain in the pGADT7 vector or downstream of the GAL4 DNA binding domain in the pGBKT7 vector. We introduced paired constructs into strain AH109 of Saccharomyces cerevisiae (Clontech, United States) and selected on SD/-Leu/-Trp (synthetic defined plate deficient in both Leu and Trp) plates. The interactions were examined on SD/-Leu/-Trp/-His/-Ade plates.

Preparation of Microsomal Proteins
Fractionation was performed basically as described previously (Li et al., 2006). Two grams of roots from 7-day-old seedlings were harvested and ground to fine powder in liquid nitrogen. Ground tissues were suspended in a homogenization buffer (50 mM Tris-HCl, 2 mM ethylene diamine tetraacetic acid (EDTA), 10 mMβ-mercaptoethanol, 250 mM sucrose, pH 7.5) and centrifuged at 8,000 × g for 15 min at 4 • C to remove debris. The supernatant was recovered, and we repeated centrifugation. The resulting supernatant was ultracentrifuged (Optima TM L-100 XP Ultracentrifuge; Beckman Coulter, United States) at 100,000 × g for 1 h at 4 • C. The pellet was surface washed with 80% cold acetone and subjected to proteomics analysis.

Label-Free Analysis
Label-free analysis was performed as described previously (Sheng et al., 2015), with modifications. The abundance of each protein in multiply samples was normalized by total intensity. Briefly, peptides were harvested by centrifugation, acidified with 1% CF3COOH, and subsequently dried with a refrigerated CentriVap concentrator (Labconco, Kansas, MO, United States). The dried peptide mixture powder from each digested sample was reconstituted with 30 µl 2 mM TEAB buffer (pH 8.5). Prior to mass spectrum (MS) analysis, samples were desalted onto an Empore C18 47-mm disk (3M) (Ishihama et al., 2006). The dried peptides were resuspended in 0.1% (v/v) formic acid solution and then analyzed with a Q Exactive mass spectrometer (Thermo Electron Finnigan, San Jose, CA, United States). The mass spectra were submitted to the Maxquant software (version1.4.1.2) for peptide identification, and searched against A. thaliana protein sequences (Tair) downloaded in 2014. The following parameters were used: carbamidomethylation of Cys was set as fixed modification, phospholation of STY, oxidation of M, and acetylation of protein N terminal were set as variable modifications, and a maximum of two missed cleavages was allowed. The false discovery rate for peptide, protein, and site identification was set to 1% (Cox et al., 2011). A total of 4,546 proteins were identified in both wild-type and mag2-1 mal-1, and 515 were differently accumulated proteins (DAPs) in the mag2-1 mal-1 double mutant. The DAPs were filtered with change ratio >1.2 or p ≤ 0.05, and 124 of the DAPs met the requirements.

Gene Ontology Enrichment Analysis
For function enrichment analysis, Gene Ontology (GO) analysis was conducted on the identified differently expressed genes (DEGs) (Ashburner et al., 2000) using online OmicShare tools. 1 First, all the DEGs were mapped to GO terms in the Gene Ontology database, 2 gene numbers were calculated for every term, and significantly enriched GO terms in the DEGs compared to genome background were defined by hypergeometric test. Calculated p-values underwent FDR correction with FDR ≤ 0.05 as the threshold. Finally, we filter out excessive terms in the three main categories [biological process (BP), MC, and cellular component (CC)].

MAIGO2 and MAG2-Like Play an Important Regulatory Role in Plant Growth and Development With Functional Redundancy
Our previous study demonstrated that MAG2 plays a crucial role in ER export (Li et al., 2006(Li et al., , 2013. MAG2 has a homologous protein, MAG2-like (MAL) (Zhao et al., 2013;Zhao and Lu, 2014). Both of them have similar gene structure ( Figure 1A) and protein structure that contain a conserved RINT-1/TIP20 domain ( Figure 1B). Tissue expression determination revealed that MAL was expressed in all tissues, with highest level in roots, followed by rosette leaves, inflorescences, and seedlings, and with lowest level in stems and siliques ( Figure 1C).
In order to analyze the function of MAL, we isolated a T-DNA insertion mutant, mal-1, in which T-DNA was inserted in the fourth exon ( Figure 1A). Northern blot and RT-PCR analysis indicated that MAL expression was depleted in mal-1 and reduced in mag2-1, but that MAG2 expression had no significant change in mal-1 ( Figure 1D and Supplementary Figure 1A). In order to investigate the regulatory function MAG2 and MAL in plant development, we crossed mal-1 with mag2-1 to generate a double mutant. We also generated TAP-MAL overexpression (MAL/OE) plants. RT-PCR and immunoblot analysis indicated higher expression levels of MAL in MAL/OE lines (Supplementary Figures 1A,B).
The germination ratio of mal-1 and mag2-1 single mutants did not have significant change compared with that of the wild type (Figures 2A,B), but green leaf ratio was lower than that of the wild type (Figures 2C,D). The germination ratio and green leaf ratio of the mal-1 mag2-1 double mutant were significantly reduced, but those of MAG2/OE and MAL/OE did not change significantly compared with those of the wild type (Figures 2A-D). The primary root length of seven-day-old seedlings of mag2-1, mal-1, mal-1 mag2-1 mutants and MAL/OE line was significantly shorter than that of wild type, especially mal-1 mag2-1 double mutant. However, there was no significant difference between wild type and MAG2/OE (Figures 2E,F). Noticeably, the lateral root (LR) number of 14-day-old seedlings of the mal-1 mag2-1double mutant was higher than that of the wild type, but there was no significant difference among the other lines (Figures 2G,H). The LR length of mal-1 mag2-1 and MAG2/OE and MAL/ OE lines were significantly longer than that of wild type, especially the double mutant was more than twice of wild type. However, there was no significant difference between the mal-1 and mag2-1 single mutants and the wild type ( Figure 2I). The aerial part and rosette leaves of 36-day-old plants of mag2-1, mal-1, and mal-1 mag2-1 were smaller than those of the wild type ( Figure 2J), especially the double mutant, while the 70-day-old plant height of all the mutants and OE lines were shorter than that of the wild type, especially the double mutant ( Figures 2K,L).
The seed number per silique of mag2-1, mal-1, and mal-1 mag2-1 was significantly less than that of the wild type, especially mal-1 mag2-1 (Figures 2M,N). The seed size of mag2-1, mal-1, and mal-1 mag2-1 was smaller than that of the wild type, but that of MAG2/OE and MAL/OE was larger than that of the wild type ( Figure 2O). Consistent with this, the thousand grain weight (TGW) of mag2-1 and mal-1 mag2-1 was significantly lower than that of the wild type, that of MAG2/OE and MAL/OE was significantly higher than that of the wild type, and that of mal-1 was similar to that of the wild type ( Figure 2P). The above results suggest that both MAG2 and MAL are involved in regulation of plant growth and development. The fact that the phenotypes of the mal-1 mag2-1 double mutant were more serious than those of the mal-1 and mag2-1 single mutants suggest that both MAG2 and MAL play important roles in plant growth and development with functional redundancy.

MAIGO2 and MAG2-Like Regulate Protein Transport With Redundancy
Our previous research clarified that MAG2 forms a complex with MIP1, MIP2, and MIP3 to regulate Golgi-to-ER retrograde transport (Li et al., 2006(Li et al., , 2013. MAG2 also interacts with ER-localized Qa-SNARE AtSYP81 and Qb-SNARE AtSec20 to promote membrane fusion (Li et al., 2013). In the mag2, mip1, mip2, and mip3 mutants, ER export of SSP precursors is blocked, which results in accumulation of proteins inside the ER and serious ER stress (Li et al., 2006(Li et al., , 2013Zhao et al., 2018;Guan et al., 2021). In order to explore the role of MAL in vesicle transport, we first performed yeast two hybrid analysis to detect the interaction between MAL and SNARE and MAG2 complex subunits. The results indicated that MAL interacted with AtSYP81, AtSec20, MIP1, and MIP2, and that MAG2 interacted with AtSYP81, AtSec20, and MIP1 ( Figure 3A), suggesting that MAL has the ability to form a complex with MIP subunits to regulate arriving vesicle tether to the ER membrane, maybe coordinately with SNAREs. Immunoblot analysis revealed that a trace amount of SSP precursors accumulated in the mal-1 seeds, and that numerous precursors accumulated in the double mutant seeds and were more than those in mag2-1 (Figures 3B,C). This suggests that MAL plays a minor role in protein transport and that it is functionally redundant with MAG2. Since abnormal accumulation of proteins in the ER lumen induces ER stress (Li et al., 2006(Li et al., , 2013Zhao et al., 2018;Guan et al., 2021), we detected the expression of ER stress markers. Western blot of BiP1/2, a common ER stress marker, indicated that their protein accumulation was significantly increased in all the mutants, especially in the mal-1 mag2-1 double mutant ( Figure 3D). Moreover, RT-PCR determination of BiP3, and ER stress-specific marker, indicated that the transcription of BiP3 also increased significantly in all the mutants ( Figure 3E). These results suggest that protein export from the ER is blocked, inducing ER stress in mal-1. To further clarify the function of MAL, we detected the protein accumulation of MAG2 complex subunits in the mal-1 mutant. As shown in Figure 3F, in mal-1, the protein levels of MAG2 and MIP1 decreased, while that of MIP2 increased; in mag2-1, the protein levels of both MIP1 and MIP2 decreased; in the mal-1 mag2-1 double mutant, MIP1 and MIP2 decreased more than in the single mutants, indicating that both MAG2 and MAL affect the stability of the MAG2 complex. The above results suggest that MAL also plays a role in protein transport, that MAG2 function is dominant, and that they are functionally redundant.

Proteomics Analysis of Microsomal Membrane Proteins in the mag2-1 mal-1 Double Mutant
To further elucidate the effects of simultaneous depletion of MAG2 and MAL on cellular activities, we performed a proteomics analysis using extracted microsomal fraction from roots of 7-day-old seedlings by label-free identification. A total of 4,546 proteins from both the wild type and mag2-1 mal-1 were identified, 515 of which were DAPs in mag2-1 mal-1 compared with those in the wild type. The DAPs were filtered by a change ratio >1.2 or p ≤ 0.05, and 124 DAPs met the requirements (Supplementary Table 1). Then, the set of 124 DAPs was subjected to GO analysis to achieve a broader functional characterization. As a result, the DAPs were classified into 40 subcategories within three main categories: 20 subcategories in BP, 13 in CC, and 7 in molecule function (MF) (Figure 4A). In total, 110 DAPs were associated with BP terms (GO:0008150), 115 DAPs were associated with CC terms (GO:0005575), and 107 DAPs were associated with MF terms (GO:0003674) (Supplementary Table 2). Among these, one DAP could be assigned to more than one category. In BP, the most enriched pathways were response to stimulus ( (Figure 4F). The functional categories of GO terms of BP, CC, and MF were shown as a diagram (Figures 4C,E,G).
These results indicate that metabolism, biosynthesis, signaling, and environmental response were affected in the mag2-1 mal-1 double mutant.

Depletion of MAIGO2 and MAG2-Like Affects Intracellular Transport
We first extracted DAPs related to vesicle trafficking (Supplementary Table 3) and restored their functional location (P) Statistics of thousand grain weight. PR, primary root; LR, lateral root; TGW, thousand grain weight. * p < 0.05, * * p < 0.01, and * * * p < 0.001. Significance was evaluated by Student's t-test using the IBM SPSS Statistics 26 software. The seeds used in this study were all newly harvested.

MAIGO2 and MAG2-Like Deficiency Influenced Protein Quality Control
The ER is crucial for maintenance of cellular homeostasis, because its functions in various cellular processes, such as folding and initial modification of secretory and transmembrane proteins. Misfolded and unfolded proteins that accumulate in the ER lumen induce ER stress. To maintain ER homeostasis, several strategies have been evolved, including unfolded protein response (UPR), ER-associated degradation (ERAD), and ER-phagy (Chen et al., 2020). As shown in Figure 5C, the protein levels of ERlocalized CDC48B ATPase (extracts unfolded/misfolded proteins from the ER lumen and membrane for targeting to proteasomes) (Wu and Rapoport, 2018) and MNS4 mannosidase (promotes BRI1-5 ubiquitylation and degradation) (Hüttner et al., 2014) in the ERAD pathway, and Sec62, an ER-phagy receptor (coordinates with ATG8e to engulf misfolded proteins into autophagosomes for vacuolar degradation) (Hu et al., 2020) were increased. These changes have probably resulted from blocking of protein export from the ER, suggesting that MAG2 and MAL are important for protein quality control. Interestingly, the quinolinic acid phosphoribosyl transferase (QPT), which is essential for pyridine nucleotide cycle and biosynthesis of alkaloid nicotine (Eads et al., 1997;Sinclair et al., 2000) decreased in mag2-1 mal-1 (Figure 5D).

MAIGO2 and MAG2-Like Affect Abundance of Regulators Related to Auxin Biosynthesis, Transport, and Signaling
Since the growth and development of the mag2-1 mal-1 double mutant were seriously affected, we then focused on auxin-related DAPs and plant phenotypic analysis. We first extracted auxinrelated DAPs (Supplementary Table 4) and restored their function (Figures 5E-G).
columella cells in primary and lateral root tips, lateral root primordium, cotyledon veins and margin, and true leaf tips in the wild type. However, in mag2-1, the DR5:GUS signal was significantly reduced, only observed in a few columella cells in primary root tips. Unexpectedly, the expression pattern of DR5:GUS in mal-1 was completely different. In mal-1 primary roots, the DR5:GUS signal increased significantly not only in QCs and columella cells but also in stele cells. Moreover, increased GUS signals were also observed in lateral root tips, lateral root primordium, cotyledon veins and margin, and true leaf tips, while in the mag2-1 mal-1 double mutant, DR5:GUS distribution was similar to that in mag2-1 ( Figure 6A). These results suggest that knockout of MAG2 and MAL affects auxin level and distribution, but that the two homolog proteins might play different regulatory roles in auxin distribution. We further detected the auxin response of the mutants and OE lines. Application of 50 nM of NAA, a synthetic auxin analog, inhibited the growth of primary roots of 7-day-old seedlings. In the wild type, root length decreased by more than 20%, while the reduction in the root length of mutants and OE lines was much less than that in the wild type (Figures 6B-D). These results suggest that MAG2 and MAL are involved in auxin response.
Then, we checked polar auxin transport (PAT). Application of 3 µM of NPA, an auxin transport inhibitor, inhibited the growth of primary roots of 7-day-old seedlings. Reduction in the root length of the mag2-1 and mal-1 single mutants and the OE lines was less than that of the wild type, while the root growth of the mag2-1 mal-1 double mutant was not sensitive to the inhibition of 3 µM NPA treatment (Figures 6B-D). These results suggest that PAT was affected in the mutants and OE lines, especially in the mag2-1 mal-1 double mutant.

PIN-FORMED Abundance Was Affected in the Mutants
To clarify the mechanisms underlying MAG2 and MAL regulation in auxin transport, we introduced the cassettes of PIN1:GUS, PIN1-GFP, PIN2:GUS, PIN2-GFP, PIN3:GUS, and PIN3-GFP into the mutants by crossing, and we observed their distribution. PIN1 is localized on cell basal side in root stele and stem vascular tissue, as well as lateral root primordium (LRP) (Omelyanchuk et al., 2016). Compared with the wild type, PIN1-GFP signals decreased significantly in stele cells in mag2-1, mal-1, and mag2-1 mal-1, especially in the mag2-1 mal-1 double mutant (Figure 7A). PIN1:GUS in stele cells of primary roots decreased significantly in mag2-1, mal-1, and mag2-1 mal-1, especially in mag2-1 mal-1. Interestingly, PIN1:GUS expression increased in QC cells in primary roots of mag2-1 and mal-1, especially mag2-1. No signal was observed in the mag2-1 mal-1 double mutant ( Figure 7B). In wild-type LRP, PIN1:GUS evenly distributed in all cells, but in mag2-1 LRP, PIN1:GUS signals increased in the basal layer. Conversely, in mal-1 LRP, PIN1:GUS signals decreased in the outer layer. No signals were detected in the mag2-1 mal-1 double mutant ( Figure 7B). PIN2 is mainly expressed in cortical and epidermal cells in root tips and is involved lateral root development (Chen et al., 1998;Zhou and Luo, 2018). In mag2-1 and mal-1, PIN2-GFP localization and abundance did not change significantly, but in the mag2-1 mal-1 double mutant, PIN2-GFP abundance likely increased ( Figure 7C). PIN2:GUS signals increased in cortical and epidermal cells in primary and lateral root tips of mag2-1 and mag2-1 mal-1 but decreased in mal-1 ( Figure 7D). In the early LR development stage, PIN2:GUS in mag2-1 tended to accumulate in basal layers compared with that in the wild type, but in mal-1, PIN2:GUS signals became weaker, whereas in the mag2-1 mal-1 double mutant, GUS signals became higher and diffused ( Figure 7D). PIN3 is distributed in root columella and stele cells , participating in primary root development and lateral root formation in early steps (Zhou and Luo, 2018). In mag2-1, mal-1, and mag2-1 mal-1, PIN3-GFP abundance in columella and stele cells reduced, especially in the mag2-1 mal-1 double mutant ( Figure 7E). PIN3:GUS was expressed in stele, columella, and LRP cells in the wild type, but almost no signal was detected in all the mutants (Figure 7F). We further determined the expression of AUX1, part of PIN and IAA genes. The results indicate that the expression of IAA1 increased and that of IAA3 reduced slightly (Supplementary Figure 1C). The alteration in abundance of PIN1, PIN2, and PIN3, and expression of IAAs in the mutants might lead to abnormal auxin transport and distribution and affect lateral root development. Combined with the proteomics results, it is suggested that auxin transport and signaling are disturbed in MAG2-and MAL-deficient mutants.

MAIGO2 and MAG2-Like Are Involved in Plant Stress Response
It was observed that in the early stage of germination, the seedlings of mag2-1, mal-1, mag2-1 mal-1, MAG2/OE, and MAL/OE lines accumulated higher levels of anthocyanins than those of the wild type (Supplementary Figure 1D). Anthocyanins are antioxidants that protect plants from growth inhibition and cell death by scavenging abiotic stress-induced ROS, thereby enabling plant adaption to abiotic stress (Naing and Kim, 2021). The higher accumulation of anthocyanins suggested loss of ROS homeostasis in the mutants and OE plants.
In order to explore the function of MAG2 and MAL in plant response to environmental stress, we performed salt, osmotic, and ABA treatments. In the 125-mM NaCl treatment, reduction of root length of the mutants was more than that of the wild type, but that of the OE lines was less than that of the wild type (Figures 8A-C). In the 200-mM mannitol treatment, reduction of root length of the mutants and OE lines was less than that of the wild type (Figures 8A-C). In the 1-µM ABA treatment, reduction of root length of the mutants was higher, and that the OE lines was less than that of the wild type (Figures 8D-F). Since ABA signaling was disrupted in mag2-1 , we checked the expression levels of ABI3 and ABI4 in mag2-1 mal-1. As shown in Figure 8G, the expression of ABI3 and ABI4 was significantly elevated in mag2-1 mal-1. All these results suggest Two independent experiments per sample, and three repeats per experiment. * p < 0.05, * * p < 0.01, and * * * p < 0.001. Significance was evaluated by Student's t-test using IBM SPSS Statistics 26. that MAG2 and MAL play important roles in regulation of plant stress response.

MAIGO2 and MAG2-Like Play Important Roles in Plant Development With Functional Redundancy and Division
Our previous study clarified that MAG2 forms a tethering complex with MIP1, MIP2, and MIP3 to regulate protein export from the ER. Deficiency of any subunit of the complex leads to the formation of a novel cell structure (we call it "mag Body"), which contains precursors of SSPs and the ER, BiP, and PDI. mag Bodies are trapped inside the ER lumen and induce severe ER stress (Li et al., 2006(Li et al., , 2013Zhao et al., 2018). In this study, we investigate the function of MAL and compared it with that of its homologs protein, MAG2. As expected, MAL also plays roles in vesicle trafficking, plant development, and environmental stress response, and it was functionally redundant with MAG2. MAL and MAG2 deficiency significantly affected the stability of the MAG2 complex (Figure 3F), indicating that MAL might form a complex with MIP proteins to regulate vesicle transport when MAG2 is deficient, or in different developmental stages or tissues.
One observation that attracted our attention was the different performance of MAL and MAG2 on auxin transport. The DR5:GUS signals were reduced significantly in mag2-1 but, conversely, were elevated substantially in mal-1, and the mag2-1 mal-1 double mutant displayed a trend similar to mag2-1 (Figure 6). Similarly, the opposite phenotypes were also observed in PIN2:GUS distribution in roots. The PIN2:GUS signals were increased significantly in mag2-1, while they were decreased markedly in mal-1 in root elongation zones. The mag2-1 mal-1 double mutant displayed a trend similar to mag2-1 ( Figure 7D). Also, PIN1:GUS expression level in LRP in mag2-1 was elevated but reduced in mal-1 (Figure 7B). These phenotypes suggest that MAL and MAG2 have a functional division in regulating auxin transport, and that their functions might be opposite: MAG2 plays a positive role, while MAL plays a negative role, and MAG2 is dominant. However, the speculation needs more evidence to be confirmed.
The mag2 and mip single mutants as well as their double mutants such as mag2-1 mip3-1 and mip2-1 mip3-1 have a distorted response to environmental stresses . The single and double mutant seeds have reduced protein qualities, germination activities, and longevity, since they have reduced content of mature SSPs, which could protect cell components and cell structures from oxidative stress during deposition. The blocking of vesicle transport in the mag2 and mip single and double mutants disturb endomembrane function and ABA signaling. The expression levels of ABI3, ABI4, and ABI5 was altered significantly compared with that of the wild type under normal and stress conditions . Consistent with these, the expression of ABI3 and ABI4 in mag2-1 mal-1 was also altered significantly ( Figure 8G), suggesting that serious blocking of ER export is bound to affect ABA signaling.

MAIGO2 and MAG2-Like Regulate Auxin Homeostasis by Controlling Golgi-to-Endoplasmic Reticulum Vesicle Trafficking
Since ER export is blocked in mag2-1, the function of ER is seriously disrupted. Numerous newly synthesized proteins are trapped inside the ER lumen and form a novel cell structure, mag Body, and subsequently induce severe ER stress (Li et al., 2006(Li et al., , 2013Zhao et al., 2018). The mag2-1 mal-1 double mutant has more serious transport defects such as more SSP precursors and higher ER stress than the mag2-1 single mutant, thus ER function disorder should be more serious. A large amount of DAPs in vesicle trafficking pathways represent the severity of the disorder (Figure 5A). The DAPs were distributed not only in the ER-Golgi COPI-and COPII-mediated pathways but also in the late secretion and recycling pathways as well as vacuole targeting pathways. This reflected the close correlation among the transport pathways. The ER is the initial point of secretory pathway and is important for ion homeostasis, quality control of newly synthesized proteins, lipid biosynthesis and transfer, and organelle communication (Borgese et al., 2006). The serious protein export jam and ER stress in mag2-1 mal-1 double mutant definitely disrupted ER homeostasis and functions, and affected the abundance of regulators of vesicle trafficking (Figure 5A), ERAD and ER-phagy pathways (Figure 5B), and lipid transfer system ( Figure 5C). As a consequence, cellular function and integrity as well as plant development were seriously affected.
Another spectacular change was the large amount of DAPs in auxin transport, signaling, and biosynthesis pathways (Figures 5E-G). Auxin homeostasis is coordinately regulated by multiple processes such as IAA biosynthesis, conjugation, transport, and signaling as. However, the controlling mechanisms of IAA homeostasis is elusive because of the complexity of combination of diverse pathways and spatiotemporal (different organs and developmental stages) and environmental factors. The two-step pathway converting tryptophan (Trp) to IAA is a highly conserved auxin biosynthetic pathway. TAA aminotransferases catalyze tryptophan to IPyA, and then YUC monooxygenases convert IPyA to IAA (Stepanova et al., 2008;Cao et al., 2019). Flower-specific YUC4.2 is the first reported ER membrane-anchored monooxygenase (Kriechbaumer et al., 2012). In Arabidopsis and maize, about half of TAA/TAR and YUC family enzymes such as TAR2, YUC3, YUC5, YUC7, YUC8, and YUC9, are localized in the ER membrane (Kriechbaumer et al., 2015(Kriechbaumer et al., , 2016. These enzymes are actively involved in auxin biosynthesis (Kriechbaumer et al., 2016;Poulet and Kriechbaumer, 2017). Moreover, about 20% of the total IAA biosynthetic activity was detected in a purified microsomal membrane fraction (Kriechbaumer et al., 2015(Kriechbaumer et al., , 2016. Thus, the ER could be considered as a platform for auxin biosynthesis. The abnormal protein accumulation inside the ER in the mag2-1 mal-1 double mutant will definitely affect the function of these auxin biosynthesis-related proteins. Endoplasmic reticulum-localized PIN5, PILS2, and PILS5 are suggested to transport auxin from the cytosol to the ER (Mravec et al., 2009;Wabnik et al., 2011). PILS2 and PILS5 are proposed to regulate auxin metabolism and signaling by increasing IAA conjugates and simultaneously decreasing nuclear auxin signaling, presumably by confining IAA in the ER (Barbez et al., 2012), whereas the pollen-specific PIN8 decreases IAA ER-compartmentation antagonistically (Dal Bosco et al., 2012;Ding et al., 2012). Therefore, these ERlocalized auxin carriers affect auxin conjugation and link IAA transport to metabolism and signaling (Barbez and Kleine-Vehn, 2013;Kriechbaumer et al., 2015). In addition, the auxindeconjugation, ILL2, IAR3, and ILR1, have been shown to localize in the ER where they are likely to produce free IAA by amidohydrolyzing IAA-amino acid conjugates (Ludwig-Müller, 2011;Sanchez Carranza et al., 2016). Considering the above clues, it is speculated that auxin conjugation could happen in the ER (Kriechbaumer et al., 2015). It is predicted that the ER functions as the main conduit for nuclear auxin uptake (Sauer and Kleine-Vehn, 2019). Given all of that, the ER serves not only as a platform for auxin biosynthesis but also as an auxin deposit and cycling hub (Friml et al., 2003). The disordered ER homeostasis and functions in the mag2-1 mal-1 double mutant might affect auxin deposition and cycling.
In mag2-1 cells, protein abundance of the phosphatase PP2A, which works antagonistically with kinase PID to regulate PIN cycling and activity, was decreased ( Figure 5A). Breaking of balance of two enzymes with opposite functions will definitely influence PIN homeostasis in the PM. As a result, the protein abundance of PIN-GFP and PIN:GUS was altered significantly (Figure 7) and subsequently affected auxin transport and response in the mag2-1 mal-1 double mutant (Figure 6).
Endoplasmic reticulum and auxin homeostasis maintenance by MAG2/MAL-mediated vesicle trafficking is essential for auxin transport and plant development, especially under stress conditions. Our study unveiled a novel perspective of membrane trafficking regulatory role in auxin homeostasis.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
LL and J-KZ conceived the project. LL and XNZ designed the experiments. XM, XMZ, HZ, ZL, YLiu, and XNZ conducted the experiments. XM, YZ, SS, YLi, and RL conducted the proteomics data analysis. HZ and LT conducted the confocal observation. LJ contributed reagents, materials, and analytical platform. LL and XM wrote the manuscript. All authors commented on the manuscript and approved the submitted version.

ACKNOWLEDGMENTS
We thank Chuanyou Li (Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing) for donating the PIN1-GFP, PIN2-GFP, and PIN3-GFP marker lines. We also thank Xugang Li (Shandong Agricultural University) for donating the PIN1:GUS, PIN2:GUS, and PIN3:GUS marker lines.