COPI-independent retrograde traffic from the Golgi to the ER depends on RAB6 and involves formation of membrane tubules instead of classical transport vesicles. COPI-dependent and COPI-independent retrograde transport appear to have distinct cargo, as anti-COPI antibodies inhibit the traffic of KDEL-containing receptors, but not that of Shiga or Shiga-like toxins, or of Golgi-resident glycosylation enzymes (White et al, 1999; Girod et al, 1999). It is not yet clear how membrane tubules formation is initiated, however cargo type and concentration, as well as lipid composition may contribute (Martinez et al, 1997; Simpson et al, 2006; de Figueiredo et al, 1998; de Figueiredo et al, 1999; reviewed in Heffernan and Simpson 2014). The presence of sn2-lysophospholipids in the Golgi membrane generates curved membranes that are thought to favour tubule formation. Lysophospholipids are generated by the activity of phopholipase A2 enzymes that hydrolyze the fatty acid at the sn-2 position; this activity is counteracted by the activity of lysophospholipid acyltransferases (LPATs). The balance of these two activities at the Golgi membrane is thought to play a role in determining whether COPI-dependent or -independent transport is favoured (de Figueiredo et al, 1998; de Figueiredo et al, 1999; Schmidt et al, 2009; reviewed in Heffernan and Simpson, 2014). Recent studies have also implicated the coiled coil homodimer Bicaudal-D (BICD) proteins in COPI-independent retrograde traffic. BICD proteins bind RAB6:GTP with their C-terminal ends and the dynein:dynactin motor complex with their N-terminal ends and in this way are thought to facilitate the recruitment of motor proteins to the RAB6 retrograde pathway (Hoogenraad et al, 2001; Matanis et al, 2002; Young et al, 2005; Januschke et al, 2007).