- Top of page
- Materials and Methods
- Supporting Information
Many integral and peripheral membrane proteins involved in signal transduction or membrane trafficking can be subjected to covalent lipid modifications, such as myristoylation, farnesylation, prenylation and S-acylation. S-acylation (also known as S-palmitoylation) is the attachment of palmitic or stearic acid to cysteine residues via a labile thioester bond (Gleason et al., 2006; Resh, 2006a,b). It is widespread in eukaryotes, often coupled with myristoylation or prenylation, and increases the lipophilicity of the modified protein, thus enhancing its membrane association. Unlike other lipid modifications, it is reversible and can accommodate regulation by extracellular signals (Tsutsumi et al., 2008; Chini & Parenti, 2009). Cycles of de- and re-acylation of peripheral membrane proteins influence their membrane-association dynamics in both mammals and yeast. For example, Ras is subjected to a dynamic acylation pathway that mediates trafficking between Golgi and plasma membrane, and the correct membrane-localized functioning of α subunits of most heterotrimeric G proteins is dependent on S-acylation (Roth et al., 2006; Greaves & Chamberlain, 2007). Integral membrane proteins such as GPCRs, ion channels and SNARE proteins are S-acylated influencing fidelity of processing and transport to specific membranes and membrane microdomains, or altering conformation such that their activity or interaction with other proteins is modified (Resh, 2006a,b). S-acylation of cysteines in transmembrane domains (TMDs), can promote lateral diffusion into thicker microdomains rich in sphingolipid and cholesterol, or can tilt the TMD. Such changes shield or expose membrane-proximal amino acids that are targets for protein–protein interaction or post-translational modifications (Greaves & Chamberlain, 2007).
S-acylation is catalysed by protein S-acyl transferases (PATs) in yeast (Lobo et al., 2002; Roth et al., 2002) and in mammals (Fukata et al., 2004; Huang et al., 2004; Keller et al., 2004). PATs are integral membrane proteins with four to six TMDs and a cytoplasmic DHHC-containing Cysteine Rich Domain (DHHC-CRD) that is essential for catalytic activity (Montoro et al., 2011). S-acyl transferases are encoded by a seven member gene family in Saccharomyces cerevisiae, 15 predicted genes in Caenorhabditis elegans, at least 23 predicted genes in Drosophila melanogaster, mammals (Tsutsumi et al., 2008) and 24 in Arabidopsis thaliana (Batistič, 2012). In addition to the DHHC domain some S-acyl transferases also contain a PDZ-binding motif, others an SH3 domain, while other members contain multiple ankyrin repeats.
In plants, understanding of S-acylation is limited. A few proteins have been shown to be S-acylated and these are involved in Ca2+ signalling, movement of potassium ions, stress signalling and pathogenesis (Hemsley & Grierson, 2008; Hemsley, 2009). A proteomic approach identified > 500 potentially palmitoylated proteins in Arabidopsis (Hemsley et al., 2013); however, so far, only one plant S-acyl transferase has been characterized, Arabidopsis TIP1. The transcriptional null mutant alleles exhibit defects in cell size control, pollen tube, root hair growth and cell polarity (Hemsley et al., 2005). Recently, a survey of the genomics and localization of the 24 Arabidopsis PATs described the ubiquitous expression profiles of most PATs, and their complex targeting patterns in cellular membrane compartments which are different from their counterparts in yeast and mammals (Batistič, 2012).
As part of an effort to determine the biological functions of Arabidopsis PATs we analysed two T-DNA insertion lines (Alonso et al., 2003) of AtPAT10 (At3g51390), investigated the activity of the protein by expression in yeast and localized C-terminally YFP tagged AtPAT10 to the Golgi, trans-Golgi network (TGN) and tonoplast. Our results demonstrate that AtPAT10 is an S-acyl transferase involved in the regulation of cell expansion, cell division, vascular development, stature, shoot branching and fertility in Arabidopsis. This greatly expands the range of events in which S-acylation is involved in Arabidopsis and reveals a Golgi and tonoplast located S-acylation mechanism that affects a range of events during growth and development.
- Top of page
- Materials and Methods
- Supporting Information
We have demonstrated that the Arabidopsis gene At3g51390 is required for the normal growth and development of Arabidopsis and encodes an S-acyltransferase, AtPAT10, that can complement the yeast S-acyltransferase AKR1 loss-of-function mutant akr1 (Fig. 1), and that it is auto-acylated (Fig. 2). Furthermore, we proved that both activities of AtPAT10 require the Cys in the conserved DHHC motif because mutation of this residue to Ala prevents complementation and auto-acylation in yeast. In addition, the loss-of-function of AtPAT10 mutant atpat10 phenotypes were rescued by ectopic over-expression of the AtPAT10 cDNA (Fig. 8), proving that the mutant phenotype is caused by the AtPAT10 gene being rendered nonfunctional. However, the AtPAT10C192A construct which carries a point mutation in the S-acyltransferase active site failed to rescue the mutant phenotype, confirming that the phenotype is caused by loss of AtPAT10 S-acyltransferase activity. Taken together, our data demonstrated that AtPAT10 is an S-acyltransferase, its enzyme activity requires the core DHHC motif, and that it is functionally independent of the other 23 Arabidopsis PATs.
Of the 24 DHHC-CRD containing proteins encoded by the Arabidopsis genome, only one, TIP1, has been functionally characterized (Hemsley et al., 2005). The fact that mutation of the DHHC motif of TIP1 to DHHA abolished its ability to rescue the yeast akr1 mutant, as well as its ability to auto-acylate in yeast (Hemsley et al., 2005), suggests that in Arabidopsis the Cys of the DHHC motif of this family of proteins may be necessary for catalytic activity, as it is known to be in yeast and mammals (Roth et al., 2002; Hou et al., 2009; Fukata & Fukata, 2010; Mitchell et al., 2010). Nevertheless, the phenotypes of the T-DNA insertion mutants in these two Arabidopsis PATs are distinct, suggesting that they function in different processes.
In order to understand where AtPAT10 is expressed within cells we transformed atpat10-1 mutant plants using 35S:AtPAT10-GFP and 35S:AtPAT10-YFP. The fact that both constructs fully rescued the mutant phenotype of atpat10-1 strongly suggests that some, if not all, of the expression we observed reflects the localization of endogenous AtPAT10. In combination with FM4-64 labelling and marker WAVE lines (Geldner et al., 2009), we demonstrate that the AtPAT10-YFP protein was predominately located in the Golgi and tonoplast in Arabidopsis leaf, root and hypocotyl cells (Figs 9, 10). FM4-64 labels a variety of cellular membrane compartments in a variety of plant cells among which are Golgi (Bolte et al., 2004), trans-Golgi network/early endosome (TGN/EE) (Geldner et al., 2003; Dettmer et al., 2006; Viotti et al., 2010), multivesicular bodies (MVB) (Otegui & Spitzer, 2008) and tonoplast (Bolte et al., 2004; Tse et al., 2004; Geldner et al., 2009). Our data obtained from crosses with marker Wave lines reveal that AtPAT10 co-localizes with the three Golgi markers, Got1p (18R), SYP32 (R22) and MEMB12 (R127) (Conchon et al., 1999; Rancour et al., 2002; Uemura et al., 2004; Chatre et al., 2005; Geldner et al., 2009). Got1P has been localized in Golgi stacks by immunogold electron microscopy (Geldner et al., 2009). Some co-localization was also found with VTI12 (R13) (Geldner et al., 2009), a tans-Golgi network/early endosome marker (Sanderfoot et al., 2001; Uemura et al., 2004). The TGN is a highly mobile organelle in plants that frequently displays independent movement and only transiently associates with the Golgi stacks (Batistic et al., 2010; Viotti et al., 2010). Therefore, our combined data from FM4-64 staining and marker WAVE lines show that AtPAT10 is located in multiple highly mobile membrane compartments that include the Golgi stack, TGN/EE and tonoplast.
In agreement with our observations, the tonoplast location of AtPAT10 has previously been reported in a proteomic study of vacuoles from Arabidopsis cell culture (Jaquinod et al., 2007). Transient expression of GFP tagged AtPAT10 driven by the Mannopine Synthase gene promoter in tobacco leaf epidermal peels demonstrated localization in both tonoplast and Golgi (Batistič, 2012). Interestingly, however, a recent study using stably transformed Arabidopsis expressing C-terminal GFP tagged AtPAT10 only showed a tonoplast localization for the protein in roots; localization in Golgi or other subcellular compartments was not observed (Zhou et al., 2013). In the study by Zhou et al., the AtPAT10 C-terminal GFP fusion was driven by a putative endogenous promoter region of c. 1.1 kb. However, using the same promoter to make a GUS fusion the authors were unable to detect GUS signal in most reproductive tissues where AtPAT10 is highly expressed and severe morphological defects are observed in mutant lines (Fig. 4 and Zhou et al., 2013). Thus, the lack of Golgi localization reported by Zhou et al. could be explained by the use of incomplete promoter regions in their constructs which could lead to suboptimal levels of PAT10 expression. Our observations add further detail to the cellular location of AtPAT10 in the AtPAT10 loss-of-function mutant of Arabidopsis rescued by stable transformation with AtPAT10-GFP and YFP constructs.
Of the 24 Arabidopsis PATs transiently overexpressed in tobacco epidermal peels, seven have been reported to be located in the Golgi and others are thought to reside in vesicles that co-localize with the plant endosomal marker AtRAB5C/RABF1/ARA6 (Ueda et al., 2001; Batistič, 2012). A dual location of ER/Golgi or Golgi/PM has been reported for several mammalian PATs and this is thought to be important for continuous cycling of proteins between membrane compartments and correct membrane localized functioning (Ohno et al., 2006; Greaves & Chamberlain, 2007). The location of AtPAT10 in multiple membrane compartments has not been reported for other PATs. This feature might be unique to AtPAT10, but confirming this will require detailed studies of the location of other AtPATs in stably transformed Arabidopsis.
Recent evidence suggests that the Golgi is the site of the core S-acylation machinery for palmitoylation of peripheral proteins in mammalian cells where acylation of proteins directs them to the secretory pathway and plasma membrane. From here they become redistributed to other cellular membranes and are ultimately de-acylated. Because most of these proteins have other lipid modifications such as myristoylation or prenylation, they are re-directed back to the Golgi where they can again be acylated and re-enter the secretory pathway (Rocks et al., 2010). Localization of AtPAT10 in the Golgi of Arabidopsis may reflect a similar mechanism in higher plants. However, Batistič (2012) has suggested that the plant cellular S-acylation machinery is functionally different compared with that of yeast and mammals because half of the AtPATs were seen to be localized to the PM in tobacco leaf peels. The yeast S-acyl transferase PFA3 is a vacuolar-localized PAT that palmitoylates the vacuolar fusion factor Vac8 and promotes vacuolar fusion (Hou et al., 2005; Smotrys et al., 2005). The fact that we observe AtPAT10-YFP in the tonoplast of pro-vacuoles and the mature vacuole may indicate a similar function for this Arabidopsis PAT.
Loss of AtPAT10 function affects vascular development (Figs 6, 7, S5). In the Arabidopsis shoot, xylem and phloem are specified from procambial cells by a complex transcriptional network comprising two types of transcription factors, HD-ZIP IIIs, and KANADIs (KANs) and the microRNAs 165/166 which are regulated by auxin and BR signalling. Mutations in many of these transcription factors affect vascular development (Caño-Delgado et al., 2010). For example, loss-of-function of one component of this transcriptional network, Ifll/Rev, causes a complete absence of interfascicular fibres (Zhong & Ye, 2001), raising the possibility that AtPAT10 may have a function in the xylary fibre developmental pathway that includes Ifl1.
The reduction in number of lignified xylem and interfascicular fibre cells in atpat10-1 suggests a possible link with lignin biosynthesis. Cinnamyl alcohol dehydrogenase-C and -D (CAD-C, CAD-D), and cinnamoyl CoA reductase (CCR) are the primary genes involved in lignin biosynthesis in the interfascicular fibre and xylem of the Arabidopsis floral stem (Sibout et al., 2005). A triple cad c cad d ccr1 mutant, ccc, has a significantly reduced level of lignin in mature stems and displays a severe dwarf phenotype and male sterility (Thévenin et al., 2011). This raises a possibility that AtPAT10 might have some role in lignin biosynthesis.
Proteins having diverse functions in plants are known to be palmitoylated, for example, the heterotrimeric G protein alpha subunit GPA1 and the gamma subunit 2 AGG2 (Adjobo-Hermans et al., 2006; Zeng et al., 2007). While the AGG2 knockout has no obvious growth defects (Trusov et al., 2008), null mutants of GPA1, gpa1-4 for example, exhibit slightly rounded leaves, reduced cell division and hypersensitivity to ABA, PAC and glucose in seed germination and early seedling development (Ullah et al., 2001; Chen et al., 2006). Some small G proteins, such as Type I and II ROPs (Rho of Plants) and AtRABF1, are also shown to be S-acylated (Lavy et al., 2002; Grebe et al., 2003; Lavy & Yalovsky, 2006; Sorek et al., 2010). The S-acylated Calcineurin B-like (CBL) proteins in Arabidopsis are involved in K+ transport, salt tolerance and ABA signalling (Cheong et al., 2003; Pandey et al., 2004; Li et al., 2006). Recently the tonoplast localized CBL2, 3 and 6 proteins have been shown to be mislocalized in transiently transformed protoplasts prepared from atpat10 cells. This could suggest that AtPAT10 is the palmitoylating enzyme for CBL2/3/6, although a direct protein–protein interaction between AtPAT10 and CBL2/3/6 could not be demonstrated due to technical difficulties (Zhou et al., 2013). However, the knockout out mutants of these CBLs individually did not exhibit an obvious phenotype (Batistic et al., 2010; Batistič et al., 2012) although double mutants of cbl2 cbl3 did share some phenotypic defects found in pat10 (Tang et al., 2012).
While the above S-acylated proteins are involved in various aspects of plant growth and development, none of them, however, displayed an identical phenotype to the atpat10 mutants. Therefore, the atpat10 phenotype may reflect the failure in S-acylation of more than one of these proteins, or of an as yet unidentified S-acylated protein or proteins.
Our identification and characterization of AtPAT10 defines a Golgi and tonoplast located Arabidopsis PAT that is functionally independent of other members of this multigene family, and demonstrates a growing importance of protein S-acylation in plants.