Autoinhibition and activation mechanisms of the eukaryotic lipid flippase Drs2p-Cdc50p

Abstract

The heterodimeric eukaryotic Drs2p-Cdc50p complex is a lipid flippase that maintains cell membrane asymmetry. The enzyme complex exists in an autoinhibited form in the absence of an activator and is specifically activated by phosphatidylinositol-4-phosphate (PI4P), although the underlying mechanisms have been unclear. Here we report the cryo-EM structures of intact Drs2p-Cdc50p isolated from S. cerevisiae in apo form and in the PI4P-activated form at 2.8 Å and 3.3 Å resolution, respectively. The structures reveal that the Drs2p C-terminus lines a long groove in the cytosolic regulatory region to inhibit the flippase activity. PIP4 binding in a cytosol-proximal membrane region triggers a 90° rotation of a cytosolic helix switch that is located just upstream of the inhibitory C-terminal peptide. The rotation of the helix switch dislodges the C-terminus from the regulatory region, activating the flippase.

Fig. 1

Cryo-EM of the apo Drs2p-Cdc50p complex. a Representative electron micrograph. b Selected reference-free 2D class averages. c Local resolution map. d Surface rendering of the 3D map colored by protein subunits and major domains and N- or C-terminal peptides

Fig. 2

Molecular architecture of the Saccharomyces cerevisiae Drs2p-Cdc50p complex. a Domain structures of the full-length Drs2p and Cdc50p. Regions not observed are in white. b A frontside and a backside view of the apo Drs2p-Cdc50p structure viewed from the membrane plane. The transmembrane domain of Drs2p is shown in green. The A, N, and P domains of Drs2p are shown in yellow, cyan, and purple, respectively. The autoinhibitory C-tail of Drs2p is shown in red. The regulatory N-tails of Drs2p and Cdc50p are shown in blue and pink, respectively. Cdc50p is shown in orange. N-glycans are shown as spheres

Fig. 3

Regulatory interactions in Drs2p-Cdc50p. a Overview of the A, N, and P domains of Drs2p showing the Drs2p C-tail (red cartoon), N-tail (blue cartoon), and the Cdc50p N-tail (pink cartoon). b Detailed interactions of the Drs2p C-tail with the A, N, and P domains of Drs2p. c Detailed interactions between the Drs2p N-tail with the A and P domains of Drs2p. d Interactions between Cdc50p N-tail and the TMD and the P domain of Drs2p

Fig. 4

A putative substrate lipid-binding site and the ATP-binding site in Drs2p. a Structural comparison between Drs2p (color) and the SERCA ATPase (PDB 3FPB; gray) in complex with cyclopiazonic acid (CPA) and ATP (which stabilizes the ATPase in the E2-P state), by aligning their respective TMDs. Regions in the top box highlight the substrate lipid-binding site and in the lower box highlight the ATP-binding site in the SERCA ATPase. b A putative substrate lipid-binding site in Drs2p, with residues lining the pocket shown as sticks. Previous mutagenesis showed that the underscored residues determine the specificity for phosphatidylserine transport in Drs2p. c–d Structural superposition of Drs2p and the SERCA ATPase by aligning their respective P domains (c) or N domains (d). The bound ATP in the SERCA ATPase is shown as spheres. The corresponding nucleotide-binding site in Drs2p is occupied by the C-terminal peptide of Drs2p. The curved red arrows in panels (c) and (d) show the domain rotations between the two structures

Fig. 5

The activated conformation of Drs2p-Cdc50p. a Structural comparison of Drs2p-Cdc50p in the apo and active conformations. The yellow and purple boxes highlight the substrate-transporting path and the C-terminal helix switch in Drs2p. b A putative substrate-transporting path in Drs2p in the apo and active conformations. c–d The putative PI4P binding site of Drs2p in the apo and active conformations. The helix switch of Drs2p rotates by about 90° toward the right in going from the apo to the active conformation. e Complementation of drs2Δ cells (selected by G418) with plasmids carrying either wild-type DRS2 (WT), drs2-W1223A, drs2-K1227A, drs2-R1228A, drs2-Y1235A, drs2-H1236A, or an empty plasmid (drs2Δ). Cells were serially diluted and incubated at 30 and 20 °C for 2 days. f The ATPase activity of the wild-type Drs2p (WT) and mutant enzyme complexes preincubated with PI4P. Each triangle represents a data point. Error bars are standard deviations estimated from three independent measurements

Fig. 6

A model for Drs2p-Cdc50p autoinhibition and activation by PI4P. a In the absence of an activation factor, the Drs2p-Cdc50p flippase exists in the membrane in an autoinhibited state. Autoinhibition is primarily achieved by the extended C-terminus wrapping around the cytosolic A, P, and N domains. Specifically, Phe-1275 and Phe-1277 occupy the ATP-binding pocket, preventing access of ATP to the catalytic site. b Activation of the flippase is initiated by the binding of PI4P in a positively charged pocket in the transmembrane region proximal to the cytosol. c The binding of PI4P leads to a 90° rotation of the helix switch away from the cytosolic domains, thereby pulling the C-terminal inhibition peptide out of the cytosolic domains, and activating the enzyme. Upon Drs2p activation, ATP and the substrate PS bind to Drs2p, but PI4P diffuses away from the flippase. The active state is likely sustained by additional factors such as Gea2p, which is known to bind to the Drs2p C-terminal inhibition loop. The three circles mark the ATP, PS, and PI4P binding sites in Drs2p