A single class of ARF GTPase activated by several pathway-specific ARF-GEFs regulates essential membrane traffic in Arabidopsis.

Abstract

In eukaryotes, GTP-bound ARF GTPases promote intracellular membrane traffic by mediating the recruitment of coat proteins, which in turn sort cargo proteins into the forming membrane vesicles. Mammals employ several classes of ARF GTPases which are activated by different ARF guanine-nucleotide exchange factors (ARF-GEFs). In contrast, flowering plants only encode evolutionarily conserved ARF1 GTPases (class I) but not the other classes II and III known from mammals, as suggested by phylogenetic analysis of ARF family members across the five major clades of eukaryotes. Instead, flowering plants express plant-specific putative ARF GTPases such as ARFA and ARFB, in addition to evolutionarily conserved ARF-LIKE (ARL) proteins. Here we show that all eight ARF-GEFs of Arabidopsis interact with the same ARF1 GTPase, whereas only a subset of post-Golgi ARF-GEFs also interacts with ARFA, as assayed by immunoprecipitation. Both ARF1 and ARFA were detected at the Golgi stacks and the trans-Golgi network (TGN) by both live-imaging with the confocal microscope and nano-gold labeling followed by EM analysis. ARFB representing another plant-specific putative ARF GTPase was detected at both the plasma membrane and the TGN. The activation-impaired form (T31N) of ARF1, but neither ARFA nor ARFB, interfered with development, although ARFA-T31N interfered, like ARF1-T31N, with the GDP-GTP exchange. Mutant plants lacking both ARFA and ARFB transcripts were viable, suggesting that ARF1 is sufficient for all essential trafficking pathways under laboratory conditions. Detailed imaging of molecular markers revealed that ARF1 mediated all known trafficking pathways whereas ARFA was not essential to any major pathway. In contrast, the hydrolysis-impaired form (Q71L) of both ARF1 and ARFA, but not ARFB, had deleterious effects on development and various trafficking pathways. However, the deleterious effects of ARFA-Q71L were abolished by ARFA-T31N inhibiting cognate ARF-GEFs, both in cis (ARFA-T31N,Q71L) and in trans (ARFA-T31N + ARFA-Q71L), suggesting indirect effects of ARFA-Q71L on ARF1-mediated trafficking. The deleterious effects of ARFA-Q71L were also suppressed by strong over-expression of ARF1, which was consistent with a subset of BIG1-4 ARF-GEFs interacting with both ARF1 and ARFA. Indeed, the SEC7 domain of BIG5 activated both ARF1 and ARFA whereas the SEC7 domain of BIG3 only activated ARF1. Furthermore, ARFA-T31N impaired root growth if ARF1-specific BIG3 was knocked out and only ARF1- and ARFA-activating BIG4 was functional. Activated ARF1 recruits different coat proteins to different endomembrane compartments, depending on its activation by different ARF-GEFs. Unlike ARF GTPases, ARF-GEFs not only localize at distinct compartments but also regulate specific trafficking pathways, suggesting that ARF-GEFs might play specific roles in traffic regulation beyond the activation of ARF1 by GDP-GTP exchange.