N-Glycosylation of TREK-1/hK2P2.1 Two-Pore-Domain Potassium (K2P) Channels

Figures

• 

  Figure 1. hTREK-1 channels harbor two putative N-glycosylation sites. (a) Two potential N-glycosylation consensus sites (yellow), consisting of an asparagine residue (N, red) followed by any amino acid except proline (x) and a serine or threonine residue (S/T, red) can be found in the extracellular part of the hTREK-1 amino acid sequence. Asparagine residues possibly modified by carbohydrates are located in positions 110 and 134. (b) Depicts a schematic membrane topology model of a hTREK-1 channel monomer consisting of two pore-forming loops (P1 and P2) surrounded by four transmembrane domains (M1–M4; top: extracellular space, bottom: intracellular space). Both potential N-glycosylation motifs are located in the extracellular M1-P1 interdomain. (c) Three-dimensional model of a hTREK-1 channel dimer, based on its crystal structure (PDB ID: 4TWK) [40,41,42]. N110 is located at the top of the overhead domain, and N134 is situated at a more lateral position. (d) A partial sequence alignment comparing hTREK-1 protein sequences of different species showing conservation of both motifs (‘*’, full conservation; ‘:’, conservative substitution; ‘.’, semiconservative substitution). (e) Immunoblot of…

• 

  Figure 2. Inhibition of N-glycosylation by tunicamycin decreases TREK-1 currents. (a) The TREK-1 currents elicited by a 500 ms depolarizing voltage step from −80 mV to +20 mV (see depicted pulse protocol) are displayed under control conditions (CTRL) and 48 h after administration of TM by incubation (left) or cytoplasmic injection (right). Representative recording of n = 6–12 cells. (b) The mean potassium currents at +20 mV (top) and the resting membrane potentials (RMPs, bottom) are provided for different time points (24 h, black bars; 48 h, white bars) after TM incubation (left; n = 3–15) or cytoplasmic injection (right; n = 3–15 cells). (c) Representative families of macroscopic hTREK-1 potassium currents recorded from Xenopus laevis oocytes by application of the depicted pulse protocol. (d) Corresponding mean step current amplitudes plotted as functions of the test pulse potential showing comparable current-voltage relationships under control conditions and after application of TM. Inserts: the data are presented relative to the maximum current amplitude measured at +60 mV. The data are given as the mean ± standard error of the mean (SEM). The zero-current levels are indicated by…

• 

  Figure 3. Molecular biological disruption of hTREK-1 N-glycosylation. (a) N-glycosylation can be disrupted either by introducing a proline residue in position +3 relative to the N-glycosylation acceptor residue (a,c,e,g,i) or by substituting the carbohydrate acceptor asparagine to glutamine (b,d,f,h,j). (a,b) Protein lysates of Xenopus laevis oocytes injected with RNA of the indicated hTREK-1-1d4 mutant constructs lacking the first, the second or both the first and second N-glycosylation motifs were subjected to immunoblotting. Changes in carbohydrate modifications are displayed as altered electrophoretic mobility. Please note that the E113P substitution resulted in incomplete disruption of N-glycosylation at N110. The immunosignals of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) are given as loading controls. (c,d) Representative current traces recorded from Xenopus laevis oocytes expressing the indicated hTREK-1 variants after 24 h (left) or 48 h (right). The currents were evoked by the application of depolarizing voltage steps from −80 mV to +20 mV. Representative current traces of n = 4–12 cells are shown. The mean current amplitudes (top) and resting membrane potentials (bottom)…

• 

  Figure 4. N-glycosylation of hTREK-1 currents, expressed in mammalian cells. To determine whether hTREK-1 channel subunits undergo N-glycosylation in mammalian cells, the relevance of the hTREK-1 N-glycosylation sites N110 and N134 was confirmed in HEK-293T cells. (a–c) Protein lysates of HEK-293T cells expressing hTREK-1-1d4 WT or mutant constructs were subjected to anti 1d4-immunoblotting under control conditions, after administration of tunicamycin (TM) or after cleavage of N-linked carbohydrates by the N-glycosidase PNGase F, as indicated by (+) or (−). Changes in carbohydrate modifications are displayed as altered electrophoretic mobility, and immunoblots of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) are provided as loading controls.

• 

  Figure 5. N-glycosylation regulates the surface expression of hTREK-1 channels. (a) EGFP-reporter coupled WT hTREK-1 channel subunits or glutamine mutants lacking either one or both N-glycosylation motifs were expressed in HeLa cells. Cell membranes stained with Alexa Fluor 594-labeled wheat germ agglutinin are depicted in red. The fluorescence signals of hTREK-1-eGFP variants are shown in green. The overlays (yellow) demonstrate colocalization of diglycosylated, monoglycosylated and nonglycosylated double-mutant channels with the cellular membrane, and show the preserved surface trafficking of deglycosylated channels. Scale bar: 10 µm. (b) Surface fractions of HEK-293T cells expressing the indicted hTREK-1 N-glycosylation-deficient variants were isolated via surface protein biotinylation, followed by streptavidin precipitation. Immunoblots of the input fractions are displayed on the left, and the mean immunosignals of the surface fractions are given on the right side (n = 3). (c) Ion channel subunit surface fractions (i.e., mean optical densities of the surface blots divided by the input fraction standardized by glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a loading control…

• 

  Figure 6. There is no evidence for O-glycosylation of hTREK-1 channels. Immunoblots of hTREK-1 are shown for control conditions, after administration of the O-glycosylation inhibitor benzyl 2-acetamido-2-deoxy-α-d-galactopyranoside (BenGal) and after incubation of the protein lysates with a mixture of O-glycosidase and neuraminidase (to yield potential O-glycosides accessible to O-glycosidase), as indicated by (+) and (−). However, no mobility shifts could be observed after treatment with BenGal or O-glycosidase, suggesting a lack of significant O-glycosylation of hTREK-1 in HEK-239T cells. The GAPDH signals are provided as loading controls.