α2-COP is involved in early secretory traffic in Arabidopsis and is required for plant growth

Highlight In Arabidopsis α2-COP is required for plant growth, Golgi structure and subcellular localization of p24δ5, with its loss of function resulting in upregulation of the COPII subunit SEC31A.


Introduction
The conventional secretory pathway in plants involves the transport of newly synthesized proteins from the endoplasmic reticulum (ER) to the Golgi apparatus and then to the cell surface or vacuole. The so-called 'early secretory pathway' involves bidirectional transport between the ER and the Golgi apparatus, which is mediated by COP (coat protein) I and COPII vesicles (Brandizzi and Barlowe, 2013). COPII vesicles are involved in protein export from the ER, whereas COPI vesicles are involved in intra-Golgi transport, where their directionality is still a matter of debate, and in retrograde transport from the Golgi to the ER. Coat proteins are involved in the selective capture of cargo proteins within the donor compartment, including the fusion machinery needed to ensure vesicle delivery, and the generation of membrane curvature to drive vesicle formation. COPI vesicles are formed at the Golgi apparatus and facilitate the retrieval of ER resident proteins from the Golgi to the ER as well as cycling of proteins between ER and the Golgi apparatus. Many type I transmembrane proteins transported by COPI vesicles bear a C-terminal dilysine-based motif, which has been proven to be recognized by the COPI coat (Jackson et al. 2012).
The key component of the COPI coat is the coatomer complex, which is essential in eukaryotes and is recruited en bloc onto Golgi membranes (Hara-Kuge et al., 1994). It is composed of seven subunits (α/β/β'/γ/δ/ε/ζ) that have been conceptually grouped into two subcomplexes, the B-(α/β′/ε) and F-subcomplex (β/δ/γ/ζ). The B-subcomplex has been proposed to function as the outer layer and the F-subcomplex as the inner layer of the vesicle coat (Jackson, 2014). However, recent structural studies revealed that the subunits are highly connected to each other, suggesting that the COPI structure does not fit with the adaptor F-subcomplex and cage B-subcomplex structure described for other coats (Dodonova et al., 2015). Following recruitment by the small GTPase ADP-ribosylation factor 1 (ARF1), in its GTP-bound conformation, and uptake of cargo, COPI polymerizes on the membrane surface in such a way that COPI coat assembly depends on both membrane and cargo binding. Several studies indicate that the β´-COP and α-COP subunits are involved in binding of cargo (i.e. proteins with a dilysine motif) through their N-terminal WD repeat domains. It has also been reported that the γ subunit interacts with ARF1 and that ζ-COP is required for the stability of γ-COP (Jackson, 2014).
Genes encoding the components of the COPI machinery have been identified in plants (Robinson et al., 2007;Gao et al., 2014;Ahn et al., 2015;Woo et al., 2015). In Arabidopsis, several isoforms of all the coatomer subunits, except for γ-COP and δ-COP subunits, have been identified. This is in contrast to mammals, where only γ-COP and ζ-COP subunits have more than one isoform, and to yeast that contains only one isoform for every subunit. Interestingly, electron tomography studies in Arabidopsis have identified two structurally distinct types of COPI vesicles (Donohoe et al., 2007;Gao et al., 2014). These different subpopulations of COPI vesicles might be formed by different coatomer isoforms. It is therefore of great interest to know whether these different COPI subunit isoforms have specific biological functions in plants by means of their functional characterization. Recently, the subcellular localization, protein interactions and physiological functions of β′-, γ-, and δ-COP subunits were investigated in Nicotiana benthamiana and tobacco BY-2 cells. It was shown that the COPI complex is involved in Golgi maintenance and cell plate formation, and that programmed cell death is induced after prolonged COPI depletion (Ahn et al., 2015).
In Arabidopsis, knockdown of ε-COP subunit isoforms has been reported to cause severe morphological changes in the Golgi apparatus and mislocalization of endomembrane proteins (EMPs) containing the KXD/E COPI interaction motif (Woo et al., 2015). Here, we have used a loss-of-function approach to characterize the Arabidopsis α2-COP isoform. Two α-COP isoforms, α1-COP (At1g62020) and α2-COP (At2g21390), have been identified in Arabidopsis and both isoforms contain an N-terminal WD40 domain that may allow them to recognize C-terminal dilysine-based motifs of COPI cargo proteins (Eugster et al., 2004;Jackson et al., 2012). We found that a loss-of-function mutant of α2-COP showed defects in growth. In addition, in the α2-cop mutant, the morphology of the Golgi apparatus was altered as was the subcellular localization of p24δ5, a protein with a dilysine motif that has been shown to cycle between the ER and the Golgi. A transcriptomic analysis of the α2-cop mutant showed upregulation of plant cell wall and endomembrane system genes, such as the COPII component SEC31A, indicating that α2-COP loss of function affects the expression of secretory pathway genes.

Recombinant plasmid production, plant transformation and transformant selection
The coding sequence of α2-COP with one human influenza hemagglutinin (HA) tag before the stop codon was synthesized commercially de novo (Geneart AG) based on the sequence of α2-COP (At2g21390). The coding sequence of α2-COP-HA was cloned into the pCHF3 vector carrying the CaMV 35S promoter (Ortiz-Masia et al., 2007) through SmaI/SalI. To complement growth defects, α2cop-3 plants were transformed with the α2-COP-HA construct via Agrobacterium using the floral deep method according to standard procedures (Clough and Bent, 1998). To estimate the number of T-DNA insertions in the transgenic plants, 40 seeds of each T1 line were plated on 1/2 x Murashige Skoog (MS) basal salts, 1% sucrose, 0.6% agar with 50 mg/l kanamycin. Only lines where the proportion of kanamycin resistant to sensitive plants in their progeny fitted a 3:1 ratio were considered. Homozygous α2-COP-HA T2 lines were identified by the same method.
For confocal studies, α2-cop-3 and wild type plants were transformed with a sialyl transferase (ST)-YFP construct (kindly provided by Dr DG Robinson, Heidelberg, Germany) and a RFP-p24δ5 construct, as described above. The RFP-p24δ5 construct was obtained by subcloning the RFP-p24δ5 coding sequence (Langhans et al., 2008) in the pCHF3 vector through KpnI/BamHI. The transformants were selected with antibiotic as above. We obtained three independent ST-YFP-and RFP-p24δ5-α2-cop-3 lines that showed the same confocal phenotypes over subsequent generations.

Reverse Transcription PCR (RT-PCR)
Total RNA was extracted from seedlings using a Qiagen RNeasy plant mini kit and 3 μg of the RNA solution were reverse transcribed using the maxima first strand cDNA synthesis kit for quantitative RT-PCR (Fermentas) according to the manufacturer's instructions. Semi-quantitative PCRs (sqPCRs) were performed on 3 μL cDNA template using the PCR Master kit (Roche). The sequences of the primers used for PCR amplifications are included in Supplementary  Table S1.
Quantitative PCR (qPCR) was performed using StepOne Plus from Applied Biosystems with SYBR Premix Ex Taq TM (Tli RNaseH Plus) from Takara Bio according to the manufacturer's protocol. Each reaction was performed in triplicate with 100 ng of the first strand cDNA and 0.9 μM primers for SEC31A or 0.3 µM of primers for all other genes in a total volume of 20 μl. Data are the mean of two biological samples. The specificity of the PCR amplification was confirmed with a heat dissociation curve from 60°C to 95°C. Relative mRNA abundance was calculated using the comparative Ct method according to Pfaffl (2004). Primers used for qPCR are listed in Supplementary Table S1.

Microarrays
Four-day-old seedlings of α2-cop-3 and wild type grown in 1/2 MS plates were used. Total RNA from four pools of seedlings were extracted using the RNeasy plant mini kit (Qiagen), and RNA integrity was tested by 2100 Bioanalyzer (Agilent). The four replicas of wild type were pooled together to generate a unique reference sample that was tested against each of the individual α2-cop-3 samples to generate four biologically independent assays, using the two-color dye-swap microarray. RNA labeling and microarray details were as described in Vera-Sirera et al. (2015). 0.5 μg of RNA per sample was amplified and labelled with the Agilent Low Input Quick Amp Labelling Kit. Agilent Arabidopsis (V4) Gene Expression 4344K Microarrays were used. Hybridization and slide washing were carried out with the Gene Expression Hybridization Kit and Gene Expression Wash Buffers. Slides were scanned in an Agilent G2565AA microarray scanner at 5 µm resolution in a dual scan for high dynamic range. Image files were analyzed with the Feature Extraction software 9.5.1. Raw microarray data (accession number GSE81049) were deposited in the Gene Expression Omnibus (GEO). Interarray analysis were performed with the GeneSpring 11.5 software. Only features for which the 'r or/and gIsWellAboveBG' parameter was 1 in at least three out of four replicas were selected. T-test analysis was carried out with Benjamin-Hochberg metrics to identify significantly expressed genes with a p-value below 0.05 after correction for multiple testing and where the expression ratio was above or below a twofold difference (Log 2 ±1). Features were converted into genes based on a BLAST analysis extracted from ftp://ftp.arabidopsis.org/Microarrays/Agilent/; oligo probes that did not correspond to any gene or to more than one gene were removed for further analysis. Gene ontology (GO) analysis at the biological process level was carried out with agriGO (http://bioinfo. cau.edu.cn/agriGO/; Du et al., 2010). Only GO terms with corrected a p-value less than or equal to 0.05 were selected.

Transmission electron microscopy
For electron microscopy, seedlings were grown on MS medium containing 1% agar and harvested after 4 days. Chemical fixation of cotyledons was performed according to Osterrieder et al. (2010). Ultrathin 70 nm sections were cut on a Microtome Leica UC6, stained with uranyl acetate and lead citrate and observed with a JEM-1010 (JEOL) transmission electron microscope. High pressure freezing was performed as described previously (Tse et al., 2004;Gao et al., 2012) and samples were analyzed in a Hitachi H-7650 transmission electron microscope.

α2-cop mutants display a dwarf phenotype
To investigate the function of the two isoforms of the α-COP subunit in Arabidopsis, T-DNA insertion mutants were identified and analyzed. A mutant of α1-COP, α1cop-1, which had the insertion in the third coding exon, was identified in the Salk collection, corresponding to stock number SALK_078465 (Fig. 1A). Although truncated transcripts were detected (see Supplementary Fig. S1), RT-PCR analysis confirmed that this mutant lacked the full length α1-COP transcript (Fig. 1B). As shown in Fig. 1C, the α1-cop-1 mutation did not compromise plant growth under standard growth conditions.
Here we focused on the characterization of the α2-cop-3 mutant for further analysis of α2-COP loss of function (Fig. 2). To confirm that the developmental defects in α2cop-3 were indeed caused by the loss of α2-COP function, we transformed these mutants with α2-COP cDNA containing a HA tag. As shown in Fig. 2D, the expression of α2-COP-HA in the α2-cop-3 mutant fully rescued its developmental defects. These results indicate that α2-COP may be required for normal plant growth and development.
Next, the expression levels of total α-COP in α1-cop and α2-cop mutants were analyzed compared to wild type, in order to investigate whether the dwarf phenotype of α2-cop was due to lower expression levels of α-COP. To this end, the expression levels of total α-COP, including isoforms α1 and α2, were analyzed by RT-PCR using a pair of primers (α125 and α123) common to α1-COP and α2-COP genes (Fig. 1, 2) that can therefore amplify the cDNA of both genes in wild type plants. However, these primers can only amplify the α2-COP cDNA fragment and not the α1-COP cDNA fragment in the α1-cop mutant due to the presence of the T-DNA insertion in the mutant. Similarly, these primers can only amplify the α1-COP cDNA fragment and not the α2-COP cDNA fragment in the α2-cop mutant. As shown in Fig. 3A-C, mRNA levels of α-COP were lower in α2-cop than in α1-cop mutant, which correlates with the growth defects in α2-cop-3. On the other hand, the mRNA levels of α1-COP and α2-COP in α2-cop-3 and α1-cop-1-1 mutants, respectively, were similar to those of the wild type (see Supplementary Fig. S3), indicating that there is no expression compensation between the two α-COP genes in the mutants. These results therefore suggest that the two α-COP isoforms are differentially expressed and it is the α2-COP isoform that contributes most to total α-COP levels. The protein levels of α-COP, including both α1 and α2 isoforms, were analyzed by western blot using an antibody against the first 10 amino acids of cow α-COP (Gerich et al., 1995) that has been previously described to recognize Arabidopsis α-COP (Contreras et al., 2004a). Cow α-COP and both Arabidopsis α-COP isoforms share the same first nine aminoacids and so this antibody should recognize both isoforms. As shown in Fig. 3D, the antibody recognized a band of aproximately 130 kDa corresponding to the molecular weight of α-COP. Using this antibody, we found that the α2-cop-3 mutant also has lower α-COP protein levels than wild type and the α1-cop mutant (Fig. 3D). No specific bands from the translation of truncated transcripts were detected in the mutants (Supplemental Fig. S1).

The loss of function of α2-COP affects p24δ5 trafficking and the integrity of the Golgi apparatus
As the α2-cop mutant has defects in growth, we aimed to study if retrograde transport in the early secretory pathway was impaired in this mutant. COPI vesicles are involved in the traffic of some transmembrane proteins that constitutively cycle between the ER and the Golgi using the COPII and COPI systems, such as members of the p24 family. p24 proteins constitute a family of type I transmembrane proteins of approximately 24 kDa present on the membranes of the early secretory pathway . We have previously shown that the C-terminal cytosolic tail of the Arabidopsis p24δ subfamily proteins has the ability to interact with ARF1 and coatomer subunits through a dilysine motif and with COPII subunits through a diaromatic motif (Contreras et al., 2004a(Contreras et al., , 2004b. Using a fluorescencetagged version of one member of the p24 family, RFP-p24δ5, we have also shown that p24δ5 localizes to the ER at steady state as a consequence of highly efficient COPI-based recycling from the Golgi apparatus and that the dilysine motif is necessary and sufficient for ER localization (Contreras et al., 2004a(Contreras et al., , 2004bLanghans et al., 2008;Montesinos et al., 2012). More recently, we have found that p24δ5 interacts with ARF1 and COPI subunits at an acidic pH, consistent with this interaction taking place in the Golgi apparatus (Montesinos et al., 2014). We therefore used p24δ5 as a  Table S1). ACT7 was used as a control with 22 PCR cycles. C. Phenotype of 4-week-old (left) and 7-day-old (right) seedlings of wild type and α2-cop-3 mutant. D. Rescue of the growth phenotype of α2-cop-3 by transformation with a HA tagged α2-COP cDNA construct. Phenotypes of 7-day-old seedlings (left) and 50-dayold plants (middle) of wild-type (Col-0), α2-cop-3 and α2-cop-3 complemented with α2-COP-HA. Western blot analysis, using a HA antibody, of the two independent lines of α2-cop-3 transformed with α2-COP-HA. model protein to study COPI-dependent retrograde Golgito-ER trafficking in the α2-cop mutant. In wild type plants, RFP-p24δ5 mostly localized to the ER network (Fig. 4), as described previously (Sancho-Andrés et al, 2016). In contrast, RFP-p24δ5 localized only partially to the ER and was mostly found in punctate structures, which often appeared in clusters, in the α2-cop-3 mutant (Fig. 4). Next, we checked the distribution of sialyl transferase-YFP (ST-YFP), a specific membrane marker for plant Golgi (Boevink et al., 1998), in the α2-cop-3 mutant. As shown in Fig. 5, ST-YFP showed the typical punctate pattern characteristic of normal Golgi stacks in wild type plants. However, in the α2-cop-3 mutant ST-YFP localized partially to the ER network and to clusters of punctate structures (Fig. 5).
To gain insight into the defects observed in the α2-cop-3 mutant at the ultrastructural level, we performed transmission electron microscopic (TEM) analysis of ultrathin sections of seedlings processed either by chemical fixation (Fig. 6A) or high pressure freezing/freeze substitution (Fig. 6B). The α2-cop mutant showed clear changes in the Golgi apparatus, which in most cases had a reduced number of cisternae per Golgi stack. In addition, the α2-cop mutant also contained many abnormal vesicle clusters around the Golgi remnants.

Transcriptomic analysis of α2-cop mutant
Comparative gene expression analyses were performed to gain insights into the molecular phenotype of the α2-cop-3 mutant. Global profiling analysis was carried out from 4-dayold seedlings, when the mutant growth phenotype starts to be visible, to detect early changes in expression. A median log 2 ratio of 1, namely a two-fold difference in expression, of the four biological replicates was used as cut-off criteria to compare the mutant with wild type plants. We identified 534 differentially expressed genes in the α2-cop-3 mutant; 353 induced (Supplementary Table S2) and 181 repressed (Supplementary Table S3). Confirmation of microarray data was carried out by RT-PCR in the α2-cop-3 mutant (Fig. 7A) as well as in the α2-cop-1 and α2-cop-2 mutants (see Supplementary Fig. S4). GO analysis was performed Fig. 4. α2-cop-3 mutant shows abnormal distribution of RFP-p24δ5. Confocal laser scanning microscopy of epidermal cells of 4.5-DAG cotyledons. RFP-p24δ5 mainly localized to the ER network in wild type plants (Col-0). In contrast, it was mostly found in punctate structures, which often appeared in clusters, in the α2-cop mutant. Scale bars, 10 µm. sqRT-PCR analysis of α-COP with α125 and α123 primers. Total RNA was isolated from 7-day-old seedlings of wild type (Col-0), α1-cop-1 and α2-cop-3 mutants. ACT7 was used as a control with 22 PCR cycles. A. Total α-COP expression in wild type and α1-cop-1 mutant. B. Total α-COP expression in wild type and α2-cop-3 mutant. C. Quantification of the experiments shown in A and B from three biological samples. Values were normalized against the α-COP fragment band intensity in wild type that was considered to be 100%. Error bars represent SEM. D. Western blot analysis of total protein extracts from 7-day-old seedlings of wild type, α1cop-1 and α2-cop-3 mutants using an N-terminal α-COP peptide antibody to detect both isoforms. 10 μg of total protein was loaded in each lane. GAPC was used as a loading control.
by agriGO (Du et al., 2010)    were 'plant-type cell wall' and 'endomembrane system' (Supplementary Tables S4, S6). GO terms significantly overrepresented in the biological process category among the upregulated genes were 'lipid transport' and 'cell wall modification', while one of the terms overrepresented in the molecular function category was 'oligopeptide transport activity' (Supplementary Table S4).
Interestingly, one of the genes highly upregulated was SEC31A, which encodes one subunit of the COPII coat. The Arabidopsis genome encodes two SEC31 isoforms, SEC31A (At1g18830) and SEC31B (At3g63460). Confirming the microarray data, RT-PCR analysis indicated that the expression of SEC31A in the α2-cop-3 mutant is more than 10 times higher than in wild type (Fig. 7B). SEC31A expression was also increased two-fold in the α1-cop-1 mutant. On the other hand, SEC31B expression was not upregulated in either α2-cop-3 or α1-cop-1 mutants. This strong upregulation of SEC31A in the α2-cop-3 mutant is consistent with COPII machinery being transcriptionally regulated by the alteration in COPI traffic. However in the α2-cop-3 mutant this upregulation of SEC31A seems to be specific for this particular COPII subunit as the expression other COPII subunit genes did not change (Fig. 7C, Supplementary Tables  S2, S3). Finally, we also analyzed the expression of SEC31A in gnl1 (SALK_091078C), a loss-of-function mutant for the ADP-ribosylation factor guanine nucleotide exchange factor (ARF-GEF) GNL1 that regulates COPI formation (Richter et al., 2007;Teh and Moore, 2007;Nakano et al., 2009;Du et al., 2013). Fig. 7D shows that the patterns of activation and repression for some genes that were used to confirm the microarray data of the α2-cop-3 mutant, were similar in the gnl1 mutant. This included increased expression of SEC31A, suggesting a correlation between the alteration in COPI function and changes in the expression of the COPII subunit SEC31A.

Discussion
In mammals and yeast there is only one isoform of the COPI subunit α-COP. In contrast, two α-COP isoforms, α1and α2-COP, have been identified in Arabidopsis. Here we have shown that a knockout T-DNA mutant of α1-COP resembled wild type plants under standard growth conditions. However, all α2-COP T-DNA mutants characterized had defects in root, stem and leaf growth. They were all fertile but short and bushy. The two α-COP isoforms contain at their N-terminal the WD40 domain, which it is required for KKXX-dependent trafficking. It cannot be ruled out that a truncated α-COP protein might be synthesized in the mutants that could account for the difference in the phenotypes between α1-COP and α2-COP mutants, although no truncated proteins were detected with an N-terminal antibody by western blot. On the other hand, as these isoforms share 93% amino acid sequence identity, the absence of growth defects in the α1-cop mutant might be explained by the relative expression levels of both α-COP isoforms. Nevertheless, it cannot be discounted that α2-COP may have specific functions that cannot be performed by α1-COP. Further studies will be required to clarify the differences between the functions of α1-COP and α2-COP. In this study we focused on characterizing the α2-cop-3 mutant. As the expression of α2-COP-HA in the α2-cop-3 mutant fully rescued its developmental defects, the results presented here indicate that α2-COP has a role in plant growth.
COPI vesicles are involved in the retrieval of ER-resident proteins from the Golgi apparatus to the ER and also in the traffic of other transmembrane proteins found in the early secretory pathway that are continuously cycling between the ER and Golgi, as is the case for p24 family proteins. Arabidopsis p24δ proteins contain in their cytosolic C-terminus both a dilysine motif in the -3, -4 position, which interacts with Fig. 7. Expression of specific genes in α1-cop-1, α2-cop-3 and gnl1 mutants. A. sqRT-PCR validation of the microarray data performed for four genes whose expression changed in the α2-cop-3 mutant. Total RNA was extracted from 4-day-old seedlings. Specific primers were used and ACT7 was used as a control. B. RT-qPCR analysis of SEC31A and SEC31B expression in α1-cop-1 and α2-cop-3 mutants. Total RNA was extracted from 7-day-old seedlings. The mRNA was analyzed by RT-qPCR with specific primers and normalized to UBQ10 expression. Results are from two biological samples and three technical replicates. mRNA levels are expressed as relative expression levels and represent fold changes of mutant/wild type. Values represent mean ± SE of the two biological samples. C. sqRT-PCR analysis of COPII subunit genes in the α2-cop-3 mutant. Total RNA was extracted from 4-day-old seedlings. Specific primers were used and ACT7 was used as a control. D. sqRT-PCR analysis of SEC31A, BIP3 and PILS4, which showed altered expression in α2-cop-3, in gnl1 (SALK_091078C). Total RNA was extracted from 4-day-old seedlings. Specific primers were used and ACT7 was used as a control. The pattern of expression of all three genes is similar in both α2cop-3 and gnl1. All specific primers are shown in Supplementary Table S1. COPI via α-COP and β´-COP subunits (Jackson et al., 2012;Eugster, 2004), and a diaromatic motif in the -7, -8 position involved in COPII binding (Contreras et al., 2004b). At steady state, p24δ5 mainly localizes to the ER as a consequence of its highly efficient COPI-based recycling from the Golgi apparatus (Langhans et al., 2008;Montesinos et al., 2012;2014;Sancho-Andrés et al. 2016). Here, we found that loss of α2-COP causes obvious defects in trafficking of RFP-p24δ5, which mostly localized to clusters of punctate structures and was only partially found in the ER network. This probably reflects the inability of p24δ5 to enter standard COPI vesicles for its Golgi to ER retrograde transport. In addition, the localization of the Golgi marker ST-YFP was also altered, which might be the result of fragmentation of the Golgi apparatus, consistent with recent reports showing that silencing of ε-COP and δ-COP in Arabidopsis and β′-COP in N. benthamiana results in disruption of the Golgi structure (Ahn et al., 2015;Woo et al., 2015). Indeed, the ultrastructural studies of the α2-cop-3 mutant revealed severe morphological changes in the Golgi apparatus. These results confirm the role of COPI in mantaining normal cellular function and organelle morphology in the plant early secretory pathway, as previously described in mammals and yeast (Guo et al., 1994;Gaynor and Emr, 1997;Styers et al., 2008).
Results from the microarray revealed upregulation of plant cell wall and endomembrane system genes. As most of these genes encoded proteins that follow or regulate the secretory pathway, this change in gene expression could be a mechanism to compensate for failures in the secretory pathway of the mutant due to the absence of α2-COP. Interestingly, one of the upregulated genes in the α2-cop-3 mutant was the SEC31A gene, which encodes one of the two COPII SEC31 isoforms of Arabidopsis. No changes in COPII subunits other than SEC31A have been detected. SEC31A shows 61% amino acid sequence identity with SEC31B and according to public microarray data (Zimmermann et al., 2004) they are expressed differently in Arabidopsis tissues, with SEC31B expression being about 10 times higher than that of SEC31A. It has been reported that SEC31B is not able to complement the secretion defect of the sec31-1 mutant in yeast (De Craene et al., 2014). In that study, SEC31A was not tested and the authors concluded that SEC31A could be the one that complements the secretion defect of the sec31-1 mutant. In mammals, two SEC31 isoforms have also been identified but their specific roles have not yet been defined. On the other hand, there is evidence that SEC31 interacts directly with SAR1, a small GTPase that controls COPII vesicle biogenesis, to promote SEC23 GTPase Activating Protein activity (Bi et al., 2007). It has been proposed that differences in the affinity for SEC31 between mammalian paralogs of SAR1 together with changes in the stimulated rate of GTP hydrolysis may together cooperate with the intrinsic flexibility of the outer cage in determining COPII vesicle size during assembly and disassembly of the coat on a growing bud (Zanetti et al., 2011). There is increasing evidence that indicates specific expression patterns in COPII subunit isoforms in Arabidopsis may reflect functional diversity (Chung et al, 2016). Since the expression of SEC31A was highly increased in the α2-cop mutant, SEC31A could compete with SEC31B for SAR1 binding resulting in changes in the process of assembly and disassembly of the coat that could adapt ER export machinery under these conditions. The induction of SEC31A might enable efficient packaging of specific cargo proteins into anterograde vesicles or simply increase the overall capacity of anterograde transport to compensate for the effects of the inhibition of retrogade transport in the α2-cop mutant. Interestingly, SEC31A is also upregulated in the unfolded protein response mediated by the inositol requiring enzyme-1 (IRE1), a response that is known to result in a specific remodeling of the secretory pathway (Nagashima et al., 2011;Song et al. 2015). Finally, we also found that SEC31A is also strongly upregulated in gnl1, a mutant of the ARF-GEF GNL1 involved in COPI assembly. These data suggest that the increase in SEC31A expression might be part of a general response to alterations in the secretory pathway.

Data Deposition
Raw microarray data. Gene Expression Omnibus (GEO). Accession number GSE81049. h t t p s : / / w w w. n c b i . n l m . n i h . g o v / g e o / q u e r y / a c c . cgi?acc=GSE81049