Abstract
Podocytes are visceral epithelial cells supporting the function of the glomerular filter. Interdigitating foot processes of podocytes enwrap the glomerular capillaries and are connected by a highly specialized cell junction, the slit diaphragm. Signal transduction at the slit diaphragm is essential for the proper function of the kidney filtration barrier. The slit diaphragm constitutes a dynamic multiprotein signaling complex that contains structural proteins, receptors, signaling adaptors, ion channels, and scaffolding proteins. Function of some of these proteins requires cholesterol attached to the multiprotein complex. Recruitment of cholesterol is achieved through the PHB domain protein podocin, a member of a novel family of lipid-binding proteins that are conserved through evolution. The finding that cholesterol interaction regulates the activity of ion channels at the glomerular filtration barrier has important implications for renal physiology and pathophysiology.
Human kidneys filter approximately 180 L/d plasma water. The glomerular filter through which ultrafiltrate passes consists of three layers: The fenestrated endothelium, the intervening glomerular basement membrane, and the slit diaphragms created by epithelial podocyte foot processes.1,2 Podocytes elaborate long, regularly spaced, interdigitated foot processes that enwrap the glomerular capillaries and form a 40-nm-wide filtration slit that is bridged by a continuous membrane-like cell junction called the slit diaphragm (Figure 1). The filtration barrier behaves as a selective sieve restricting the passage of macromolecules on the basis of their size, shape, and charge. A number of genes encoding for proteins localized to the slit diaphragm of podocytes have been identified in the past several years, and their mutations explain a growing number of genetic causes leading to hereditary nephrotic syndrome.3–10 Identification of these genes caused a lot of excitement that resulted in a great deal of focus on the critical role of slit diaphragm proteins in mediating signal transduction in the podocyte.11–14 These observations demonstrate that slit diaphragm signaling regulates podocyte cell differentiation, cytoskeletal dynamics, protein expression, and cell survival.15–22 Much insight into the signaling function of the slit diaphragm protein complex resulted from work on podocin, the slit diaphragm protein that is mutated in genetic forms of steroid-resistant nephrotic syndrome. Additional findings also suggest that slit diaphragm proteins form local lipid–protein supercomplexes involving an intriguing role for cholesterol in modulating cell signaling at the filtration slit.23
The glomerular filtration barrier consists of three layers: Fenestrated endothelial cells, the glomerular basement membrane, and processes of podocytes, which are connected through the slit diaphragm protein complex. (B, blood side; U, urinary space; F, foot process; SD, slit diaphragm; BM, basement membrane; E, endothelial cell.
PODOCIN AND CHOLESTEROL–AN AFFAIR TO REMEMBER1
The most common genetic form of hereditary nephrotic syndrome is a steroid-resistant disease caused by mutations in NPHS2, the gene encoding for podocin.3 Podocin localizes to the slit diaphragm of podocytes and interacts with the transmembrane adhesion protein nephrin. This interaction is required for efficient signaling through nephrin and its associated proteins.16 Podocin belongs to a large family of evolutionarily conserved membrane-associated proteins of previously unknown function.24,25 The PHB domain proteins constitute a family of more than 1300 proteins, all of which share an approximately 150–amino acid domain similar to that in the mitochondrial protein prohibitin (PHB domain24 or SPFH domain26). More than 360 of these proteins have been identified in animals, many of which have an N-terminal hydrophobic region, which places them on the inner leaflet of the lipid bilayer. On the basis of biochemical fractionation in detergent-resistant membrane preparations, a number of the mammalian PHB domain proteins co-fractionate with lipid rafts of the plasma and intracellular membranes24,26; however, until recently, the function of PHB domain proteins at the molecular level has remained unclear.
This situation has changed with the observation that podocin and related proteins can directly bind and recruit cholesterol.25 Binding of cholesterol requires the PHB domain and an adjacent small hydrophobic region of the protein that results in recruitment of cholesterol to the slit diaphragm multiprotein complex. Podocin does this by forming multimers as a high molecular weight protein complex that recruits cholesterol into this complex.18,23,27 Thus, by directly binding cholesterol on the membrane, multimerizing, and recruiting associated proteins, podocin contributes to the formation of a megadalton protein–lipid supercomplex.
Is lipid interaction important for the function of this complex? Cholesterol is not required for podocin multimerization23; formation of the megadalton complex is not affected by cholesterol depletion, and podocin mutants deficient in cholesterol binding are able to aggregate as large complexes; however, in vivo assays revealed that cholesterol binding is essential for the function of podocin-associated ion channels. Podocin binds and regulates the transient receptor potential (TRP) channel TRPC6 at the slit diaphragm.23 Consequently, mutations in either podocin or TRPC6 result in severe proteinuric kidney disease.3,6,8 The in vivo importance of cholesterol interactions is demonstrated by studies that introduced the nematode Caenorhabditis elegans as model organism to test the function of podocin-related proteins.23
The closest homologue of podocin in C. elegans is MEC-2.28,29 Podocin and MEC-2 share a similar structure with a central membrane-close hydrophobic region, amino and carboxy terminal tails facing the cytoplasm, and PHB domains that are 50% identical and 80% similar. Like podocin, MEC-2 is part of a multiprotein channel complex with the degenerin/epithelial Na+ channel proteins MEC-4 and MEC-10 that transduces gentle touch (Figure 2). In touch receptor neurons, the channel complex is localized to regular puncta along the neuronal process; MEC-2 regulates the MEC-4/MEC-10 ion channel in these puncta. C. elegans genes are also amenable to genetic manipulation. Moreover, MEC-2 function can be tested in vivo by analyzing responses of the worm to gentle touch with an eyelash attached to a toothpick; the worm responds by changing the direction of movement.30 The molecular function of podocin/MEC-2 in worms is amenable to testing with these assays: MEC-2, similar to podocin, binds and interacts with cholesterol. Mutational analysis revealed that cholesterol binding to MEC-2 is essential for the function of the MEC-2 ion channel complex but does not affect complex assembly or targeting. Touch sensation requires binding of cholesterol to the MEC-2 complex in vivo,23 and mutant worms expressing a MEC-2 protein that selectively lost the ability to bind and recruit cholesterol do not respond to touch. A special property of the nematode is its dietary requirement for cholesterol. C. elegans cannot synthesize sterols that can be used experimentally in cholesterol depletion assays. Feeding the worm cholesterol-free diets results in the loss of touch sensation. Similarly, cholesterol interaction of podocin is required for the podocin-mediated augmentation of TRPC6 current in oocytes and in cultured cells. In other words, MEC-2 and podocin bind and recruit cholesterol to organize the lipid microenvironment of the associated ion channel complexes that is essential for their function.25
Podocin and MEC-2 are associated with similar proteins.
Cholesterol is a major component of mammalian cell membranes, where it changes the physicochemical properties of the lipid bilayer; however, C. elegans does not use sterols as a structural component of the membrane.31 This argues against the need for cholesterol as a structural membrane component but rather favors a primary role for sterols in the function of protein–lipid ion channel supercomplexes. The classic lipid raft hypothesis postulates lipid microdomains rich in cholesterol and glycosphingolipids, distinct liquid-ordered phases in the lipid bilayer, dispersed in a liquid-disordered matrix of unsaturated glycerolipids.32,33 Although novel data do not argue against the existence of these rafts, one does not have to invoke this classic lipid raft model to explain new findings: Multiprotein complexes based on PHB domain–mediated homophilic interactions may recruit membrane cholesterol into the vicinity of the complex, adding a sterol-rich microsurrounding to associated proteins without the need for preformed rafts from the membrane. In fact, podocin expression in cultured mammalian cells results in the increased formation of detergent-resistant cholesterol-rich membrane fractions (T.B., unpublished data).
PALMITOYLATION REGULATES INTERACTION WITH CHOLESTEROL
If we now understand that lipid–protein interactions are important for regulating ion channel activity and the function of the slit diaphragm protein supercomplex, then the immediate question that arises is whether cholesterol interaction with the complex is regulated dynamically. Indeed, this seems to be the case. Another level of complexity is added of course by the fact that cholesterol interaction is regulated through palmitoylation of podocin and MEC-2. Palmitoylation is a reversible posttranslational modification of proteins and refers to the addition of palmitate (a 16-carbon fatty acid) to the side chain of an internal cysteine residue (Figure 3). In the case of the most common S-palmitoylation, palmitate is attached to the protein through a reversible thioesther linkage.34 Many palmitoylated proteins undergo cycles of palmitoylation and depalmitoylation, which are tightly regulated: Palmitoylation is catalyzed by protein acyltransferases, and depalmitoylation requires acylprotein thioesterases.35 The functional consequences of palmitoylation are diverse. Palmitoylation facilitates the association of proteins with membranes, regulates protein stability, and more recently identifies a targeting motif.36,37
(A) Palmitoylation, the addition of a 16-carbon fatty acid to the side chain of a cysteine residue, is a reversible posttranslational modification. (B) Cholesterol interaction of podocin and MEC-2 requires palmitoylation of the PHB domain (blue) and a hydrophobic region (yellow) adjacent to the PHB domain.
Podocin and MEC-2 are palmitoylated at two conserved cysteine residues that regulate their interaction with cholesterol.23 Mutation of the two cysteine residues to alanine not only abrogates palmitoylation but also dramatically decreases the capacity of cholesterol binding without affecting targeting to the plasma membrane or the formation of high molecular weight supercomplexes.23 Thus, dynamic changes in the palmitoylation state of PHB proteins determine the amount of cholesterol recruited to the PHB-protein–based signaling complex. Palmitoylation was previously regarded as a posttranslational modification for targeting of proteins into preformed lipid rafts.38 That palmitoylation regulates lipid-binding properties of proteins adds a new twist to this model; multiprotein complexes based on PHB domain–mediated homophilic interactions dynamically recruit membrane cholesterol and other lipids into the vicinity of the complex to add a sterol-rich microsurrounding to associated proteins, a process that is regulated dynamically through palmitoylation.
Regulation of the lipid-binding properties of the podocin-based protein–lipid supercomplex may be even more complex. however. Podocin and MEC-2 are thought to decorate the inner leaflet of the plasma membrane with a hydrophobic region at the leaflet and amino and carboxy terminal tails facing the cytoplasm. Thus, palmitoylation regulates cholesterol interaction at the inner leaflet of the plasma membrane. The C. elegans MEC-2 mechanotransduction complex contains four other proteins, and the same may be true for podocin: The degenerin/epithelial Na+ channel proteins MEC-4 and MEC-10, which are thought to form the pore of the channel (TRPC6 ion channels in the case of podocin), the paraoxonase-like protein MEC-6, and UNC-24.28,30,39,40 UNC-24 and MEC-6 may also affect the binding or metabolism of lipids associated with the MEC-4/MEC-10 channel. UNC-24 is a PHB domain protein as well, but, unlike MEC-2, it has an additional domain (sterol carrier protein domain 2) that is similar to regions of nonspecific lipid transport proteins.41 Vertebrates have similar two-domain SLP-1 proteins. Nonspecific lipid transport proteins serve as intracellular carriers of cholesterol and other sterols, so the association of a similar domain with a cholesterol-binding PHB domain is suggestive that the two domains shuttle cholesterol and other sterols into the plasma membrane.40,42 In contrast to MEC-2 and UNC-24, MEC-6 has a single membrane-spanning domain that puts most of the protein on the extracellular side of the membrane.39 The similarity of MEC-6 with paraoxonases may indicate that it, too, affects the cholesterol content of the membrane, albeit at the outer leaflet of the bilayer, because two of the three vertebrate paraoxonases are secreted and associated with cholesterol-containing HDL particles. The third paraoxonase, PON-2, is, like MEC-6, a widely expressed membrane protein.43 It is also expressed in podocytes. We speculate that MEC-6 and, by analogy, PON-2 are involved in modifying or maintaining associated lipids on the external side of the lipid bilayer, but how these proteins contribute to the long-term marriage between podocin and cholesterol at the slit diaphragm of podocytes is unclear.
MECHANOSENSATION IN THE KIDNEY
Interaction with and recruitment of cholesterol into the slit diaphragm protein complex have been suggested for several years,18,27,32 but what is the critical physiologic function? Podocin is part of a multiprotein channel complex with the TRPC6 ion channel, the transmembrane proteins neph1 and nephrin, and a number of other proteins that provide the tight link to underlying cytoskeleton. The protein composition of the slit diaphragm protein complex has striking similarities to the MEC-2 touch receptor (Figure 2). Thus, it is tempting to speculate that podocin-associated ion channels serve a similar mechanosensory function25 (Figure 4). Indeed, the TRPC6 channel is mutated in genetic forms of proteinuric kidney disease6,8 and serves as a mechanosensory ion channel.44 Hypo-osmotic and pressure-induced stretch activates TRPC6 independent of second messenger signaling and likely occurs through direct sensing of membrane stretch. Experimentally, channel activation is blocked by the tarantula peptide GsMTx-4, a peptide known to inhibit specifically mechanosensitive ion channels by inserting in the outer membrane leaflet and modifying boundary lipids that are crucial for channel exposure and opening.44 Whether this toxin directly interferes with cholesterol binding to channel-associated proteins such as MEC-2 and podocin is unclear, but why is cholesterol required for stretch activation of these channels? Cholesterol, probably together with other lipids including glycosphingolipids, modulates structural and elastic properties of the bilayer; cholesterol decreases the area per phospholipid molecule and increases the thickness of the membrane bilayer.45 Furthermore, cholesterol recruited through podocin may decrease the elasticity of the plasma membrane, resulting in a more rigid membrane directly adjacent to the channel complex.46,47
The similarity of the podocin-associated proteins to the well-defined MEC-2 touch channel complex suggests a similar role in mediating mechanosensation at the slit diaphragm of podocytes.
Taken together, it is tempting to conclude that the slit diaphragm–ion channel supercomplex is involved in mediating pressure sensation in podocytes. Podocyte secondary processes contain a highly regulated, actin-based cytoskeleton that consists of a cortical actin meshwork and highly ordered contractile central actin cables.48 Although still highly speculative, stretch-induced calcium influx at the mechanosensor of the slit diaphragm might regulate these contractile actin cables, leading to changes in foot process morphology and modulation of the filtration area and permeability of the kidney filter.
DISCLOSURES
None.
Acknowledgments
This work was supported by the Deutsche Forschungsgemeinschaft (DFG BE2212 to T.B.) and the Center for Molecular Medicine (to T.B.).
We apologize to colleagues whose work was not cited because of length restrictions.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
↵1 The 1957 movie of Cary Grant and Deborah Kerr.
- Copyright © 2009 by the American Society of Nephrology