Tetrameric assembly of CHIP28 water channels in liposomes and cell membranes: a freeze-fracture study.

Channel forming integral protein of 28 kD (CHIP28) functions as a water channel in erythrocytes, kidney proximal tubule and thin descending limb of Henle. CHIP28 morphology was examined by freeze-fracture EM in proteoliposomes reconstituted with purified CHIP28, CHO cells stably transfected with CHIP28k cDNA, and rat kidney tubules. Liposomes reconstituted with HPLC-purified CHIP28 from human erythrocytes had a high osmotic water permeability (Pf0.04 cm/s) that was inhibited by HgCl2. Freeze-fracture replicas showed a fairly uniform set of intramembrane particles (IMPs); no IMPs were observed in liposomes without incorporated protein. By rotary shadowing, the IMPs had a diameter of 8.5 +/- 1.3 nm (mean +/- SD); many IMPs consisted of a distinct arrangement of four smaller subunits surrounding a central depression. IMPs of similar size and appearance were seen on the P-face of plasma membranes from CHIP28k-transfected (but not mock-transfected) CHO cells, rat thin descending limb (TDL) of Henle, and S3 segment of proximal straight tubules. A distinctive network of complementary IMP imprints was observed on the E-face of CHIP28-containing plasma membranes. The densities of IMPs in the size range of CHIP28 IMPs, determined by non-linear regression, were (in IMPs/microns 2): 2,494 in CHO cells, 5,785 in TDL, and 1,928 in proximal straight tubules; predicted Pf, based on the CHIP28 single channel water permeability of 3.6 x 10(-14) cm3/S (10 degrees C), was in good agreement with measured Pf of 0.027 cm/S, 0.075 cm/S, and 0.031 cm/S, respectively, in these cell types. Assuming that each CHIP28 monomer is a right cylindrical pore of length 5 nm and density 1.3 g/cm3, the monomer diameter would be 3.2 nm; a symmetrical arrangement of four cylinders would have a greatest diameter of 7.2 nm, which after correction for the thickness of platinum deposit, is similar to the measured IMP diameter of approximately 8.5 nm. These results provide a morphological signature for CHIP28 water channels and evidence for a tetrameric assembly of CHIP28 monomers in reconstituted proteoliposomes and cell membranes.

Activation of epithelial Na+ channels by protein kinase A requires actin filaments.

We have recently demonstrated a novel role for "short" actin filaments, a distinct species of polymerized actin different from either monomeric (G-actin) or long actin filaments (F-actin), in the activation of epithelial Na+ channels. In the present study, the role of actin in the activation of apical Na+ channels by the adenosine 3',5'-cyclic monophosphate-dependent protein kinase A (PKA) was investigated by patch-clamp techniques in A6 epithelial cells. In excised inside-out patches, addition of deoxyribonuclease I, which prevents actin polymerization, inhibited Na+ channel activation mediated by PKA. Disruption of endogenous actin filament organization with cytochalasin D for at least 1 h prevented the PKA-mediated activation of Na+ channels but not activation following the addition of actin to the cytosolic side of the patch. To assess the role of PKA on actin filament organization, actin was used as a substrate for the specific phosphorylation by the PKA. Actin was phosphorylated by PKA with an equilibrium stoichiometry of 2:1 mol PO4-actin monomer. Actin was phosphorylated in its monomeric form, but only poorly once polymerized. Furthermore, phosphorylated actin reduced the rate of actin polymerization. Thus actin allowed to polymerize for at least 1 h in the presence of PKA and ATP to obtain phosphorylated actin filaments induced Na+ channel activity in excised inside-out patches, in contrast to actin polymerized either in the absence of PKA or in the presence of PKA plus a PKA inhibitor (nonphosphorylated actin filaments). This was also confirmed by using purified phosphorylated G-actin incubated in a polymerizing buffer for at least 1 h at 37 degrees C. These data suggest that the form of actin required for Na+ channel activation (i.e., "short" actin filaments) may be favored by the phosphorylation of G-actin and may thus mediate or facilitate the activation of Na+ channels by PKA.

A 28 kDa sarcolemmal antigen in kidney principal cell basolateral membranes: relationship to orthogonal arrays and MIP26.

Two recently cloned water channels, CHIP28 and WCH-CD, are homologous to MIP26, an integral membrane channel-forming protein found in lens fiber plasma membranes. CHIP28 is found in basolateral and apical plasma membranes of kidney proximal tubules and thin descending limbs of Henle, whereas WCH-CD is apically located in collecting duct principal cells. So far, the putative water channel that may be responsible for the high constitutive permeability of principal cell basolateral membranes has not been identified. Interestingly, freeze-fracture electron microscopy has shown that characteristic orthogonal arrays of intramembrane particles (OAPs) are found on the basolateral plasma membranes of collecting duct principal cells, and that morphologically identical OAPs present in lens fiber cell plasma membranes contain the protein MIP26. Similar OAPs have also been detected on plasma membranes of other cell types including gastric parietal cells, astroglial cells and skeletal muscle fibers. By indirect immunofluorescence, western blotting and northern blotting, MIP26 was found only in lens fibers. In addition, functional studies on reconstituted and oocyte-expressed MIP26 excluded the possibility that MIP26 might be a basolateral water channel in the kidney. However, a polyclonal antibody raised against skeletal muscle sarcolemmal vesicles, which are enriched in OAPs, produced an intense staining of principal cell basolateral plasma membranes in kidney collecting duct and immunoprecipitated a 28 kDa protein from kidney papilla. The immunoprecipitated protein from papilla was not recognized by anti-CHIP28 or anti-MIP26 antibodies, indicating that principal cell basolateral membranes contain a novel member of the CHIP/MIP family. Because this antibody also stained brain astrocyte end feet, which are enriched in OAPs, it is possible that the 28 kDa protein is related to these structures. We conclude that OAPs probably contain related but distinct proteins that may have different membrane channel functions in different cell types.