Figure S1: ARFB is targeted to the plasma membrane. To ensure further the plasma membrane localization of ARFB, we stained for 2–5 min live cells expressing ARFB:YFP with the dye FM4-64, which first labels the plasma membrane and then endocytic compartments over time (35), and with the DNA marker DAPI (36). The short time of incubation of cells with FM4-64 would ensure a predominant labeling of the plasma membrane of tobacco leaf epidermal cells. Simultaneous confocal imaging showed the plasma membrane (empty arrowheads) and chloroplasts (arrows) in the confocal imaging channel for FM4-64, while autofluorescence of the cell wall (empty arrows) as well as nucleus labeling (arrow) because of the dye were visible in the confocal imaging channel for DAPI. The cell perimeter staining of ARFB overlapped that of FM4-64 just below the cell wall (empty arrowheads in ARFB:YFP imaging channel). Importantly, no ARFB:YFP was found to concentrate at the tonoplast that is separated from the plasma membrane by the cytosol (full arrowhead in ARFB:YFP imaging channel). Scale bar = 5 μm.

Figure S2: Plasma membrane binding of ARFB depends on an intact myristoylation motif, is independent of the activation status of the GTPase and is insensitive to BFA. As the localization of ARFB had not previously been described, we wanted to explore the conditions that regulate the distribution of the GTPase to the plasma membrane. ARF proteins have a conserved glycine residue in position 2 that is part of the consensus for myristoylation (37,38). Myristoylation of this glycine residue is important for the association of ARF GTPases to membranes (39,40). To investigate whether this was also the case for ARFB, we mutated the glycine into alanine and fused the resulting mutant ARFB to YFP (ARFBG2A:YFP). A) Expression of ARFBG2Azlls showed a typical cytosolic fluorescence distribution (41), indicating that myristoylation is necessary for ARFB:YFP to bind to a membrane. We also wanted to explore whether the binding of ARFB to a membrane would depend on activation of the GTPase. For example, it has been shown that the active form of ARF1 is capable of binding to membranes, while the inactive form resides in the cytosol (8,9,19). By comparing the sequence of ARFB with those of Arabidopsis ARF1 and human ARF6, we identified T31 and Q71 that correspond to the conserved amino acid residues that have been used to generate mutants of ARF1 and ARF6 and are known to mimic the inactive (GDP) and active (GTP) status of the GTPases, respectively (20). In particular, the inactive GTPase mutants appear to have preferential affinity for GDP compared with the wild-type proteins, while the active mutants show a slow rate of GTP hydrolysis (11,20,39). To determine the subcellular localizations of the putative inactive and active forms of ARFB, we generated point mutations of the relevant residues (T31N and Q71L) and fused the resulting ARFB mutant sequences to YFP. Expression of these constructs in tobacco leaf epidermal cells showed that both were targeted to the plasma membrane (B and C), similar to wild-type ARFB:YFP (compare with Figure 1 and Figure S1). Insets on panels displaying ARFBQ71L and ARFBT31N demonstrate the plasma membrane in a cortical plane of focus. Scale bars = 5 μm. To ensure that ARFBT31N:YFP and ARFBQ71L:YFP indeed mimicked the inactive status of the GTP-bound GTPase, we tested the ability of the mutants to hydrolyze GTP by performing a GTPase activity assay on recombinant GST fusions of ARFBT31N and ARFBQ71L, and we compared the ability of these protein to hydrolyze GTP with wild-type ARFB. We anticipated that both mutations would result in a reduced ability of the GTPases to hydrolyze GTP. Therefore, equal amounts of purified recombinant GST:ARFB, GST:ARFBT31N and GST:ARFBQ71L were incubated with GTP. The GTPase activity of the proteins was then estimated as a measurement of the inorganic phosphate released as recorded by absorbance at 650 nm (see also Materials and Methods section). The activity measured with purified GST alone was used as background and subtracted from the activities of the GTPases. The activity of each mutant is presented as a ratio relative to its respective wild-type activity. The activity of ARFBT31N and GST:ARFBQ71L was clearly reduced in comparison to wild-type ARFB (D). These data support that the subcellular distribution of ARFBT31N:YFP is that of an inactive GTPase and that the exchange of GDP for GTP onto ARFB takes place in the plasma membrane, as it occurs for the human ARF6 (11). Finally, to further characterize the requirements of ARFB binding to the plasma membrane, we tested if the localization of this protein was dependent on a mechanism related to BFA-sensitive GEFs, as is the case for ARF1 (42). Therefore, we treated leaf epidermal cells coexpressing ARFB:YFP and ST:GFP (E) with BFA and followed the distribution of both markers with a confocal microscope. After 1 h of BFA treatment, we found a redistribution of ST:GFP fluorescence in the ER (F, inset), as previously reported (8,18); however, the distribution of ARFB:YFP was largely unchanged in comparison to the control (E). Insets in panels E and F show the cortex of the same cells depicted in the panels with or without BFA treatment, respectively, demonstrating that the BFA did have an effect in the cell but did not have an effect on the localization of ARFB. This indicates that the distribution of ARFB at the plasma membrane is not sensitive to BFA. Interestingly, a resistance to the inhibitory effect of BFA on the GEF responsible for activation of ARF6 has been shown also in mammalian cells (43). Scale bars = 5 μm.

Supplemental materials are available as part of the online article at

TRA_671_sm_FigureS1.tif992KSupporting info item
TRA_671_sm_FigureS2.tif3794KSupporting info item
TRA_figures1.tif992KSupporting info item
TRA_figures2.tif3794KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.