In recent years, there has been an explosion of discovery about proteins that function to organize components of the signal transduction machinery with their effectors at specific subcellular locales. Many of these molecular scaffolds control the assembly, trafficking, subcellular location and activity of epithelial transport proteins and their regulators, and, thus, are critical determinants of epithelial transport modulation. Here, the review the present state of knowledge about the two major classes of molecular scaffolds, PDZ-proteins and AKAPs, and discuss their role in renal epithelial transport.
Keywords
PDZ protein; AKAP; NHERF; Shank; Dystrophin; Lin-7; CASK
PDZ-Proteins
PDZ domains (also known as DHR domains or GLGF repeats) are ~90 amino acid, protein–protein interaction modules that bind short amino-acid motifs (4–5 residues) generally found at the extreme COOH-terminus of target proteins. More rarely, PDZ domains recognize internal sequences that mimic the COOH-terminal binding motif. The term PDZ is derived from the names of the three proteins that the structure was originally identified from (PSD 95, a post synaptic density protein), Dlg (Dropsophila Disc large tumor suppressor), and ZO-1 (zona occludens, the tight junction protein). Since its discovery as a region of sequence homology in these few proteins, the PDZ domain has become recognized as one of the most common interaction modules. The human genome contains over 250 PDZ domains in nearly 100 human proteins. The structure is evolutionarily conserved, emerging largely in metazoans, perhaps to accommodate the increased signaling needs of multicellular organisms.
PDZ domain containing proteins usually possess multiple protein–protein recognition modules. Because the domains act independently and allow concurrent recruitment of different binding targets, PDZ proteins function as molecular scaffolds. Indeed, PDZ proteins facilitate multi-protein complex formation, and organize expression of target proteins on specific membrane domains for a wide range of physiological processes. A growing body of work has strongly implicated PDZ proteins in targeting and clustering various receptors, channels, transporters, and signal transduction elements at specific plasma membrane domains in different cell types, including neurons, muscle, and the visual system. PDZ proteins play especially important roles in epithelial transport processes.
Classes of PDZ Domains
PDZ domains have been traditionally divided into three different classes, categorized by the nature of their ligands. The different ligand classes are distinguished by differences in the binding residues found at the extreme COOH of target proteins ( Figure 14.1 ). Type I domains recognize the sequence, X-S/T-X-Φ* (where X=any amino acid; Φ=hydrophobic amino acid; *=COOH terminus). Type II domains bind to ligands with the sequence X-Φ-X-Φ*. Type III domains interact with X-D/E-X-Φ* sequences. Binding specificity within each domain class can be conferred by the variant (X) residues, as well as residues outside the canonical binding motif, especially at the -3 and -4 positions (where 0 position is the C-terminal residue). Moreover, a few PDZ domains do not fall into any of these specific classes.
Based on large-scale proteomic analysis of PDZ-ligand interactions, it has been suggested the traditional three-class definition be extended to include 16 distinct binding classes. Such a classification has been proposed to predict specific interaction partners of known PDZ domains with greater fidelity than the traditional scheme.
Structural Basis for PDZ Interaction
In recent years, the structures of over 20 different PDZ domains have been solved at atomic resolution. Like many protein–protein recognition modules, PDZ domains are small globular structures. Comprised of six β-strands (βA-βF) and two alpha helices (αA and αB), PDZ domains fold into a six stranded beta sandwich ( Figure 14.1 ). The peptide ligand inserts into a binding cleft, created by the βB strand and the αB helix, effectively forming an additional antiparallel beta strand. An extensive network of hydrogen bonds and hydrophobic interactions stabilizes binding of the peptide. For instance, the conserved glycine-leucine-glycine-phenylalanine-alanine (GLGF) motif contained within a βA-βB linker provides a cradle of main chain amides, and confers recognition of the terminal carboxylate group of the peptide. A hydrophobic pocket accommodates the hydrophobic COOH-terminal residue, thereby accounting for preferential interaction with proteins ending with a hydrophobic residue (the so-called P0 position).
Binding specificity among the different binding classes is determined partly by an interaction between the P-2 residue of the target protein and the first residue of the PDZ domain αB-helix. In Class I PDZ domains, a conserved histidine residue forms hydrogen bonds with the invariant P-2 serine or threonine residue in the target protein. In class II PDZ domains, this position of the PDZ domain and the P-2 residue of the target protein are usually occupied by a hydrophobic amino acid.
Binding specificity within each domain class is also observed. At least three factors account for this. First, unique residues within or adjacent to the peptide-binding groove in the PDZ domain can interact with the target at sites other than the P-2 and P0 residues. For example, the side chain of the P-1 target protein residue usually points away from the invariant interaction surface but, in some cases, it can bond with residues that are distinct to a particular PDZ domain. Likewise, the P-3 side chain can make contact with unique residues in the interaction groove. Sites proximal to the archetypal, four amino acid-binding motif can also interact with regions outside the canonical-binding site, and thereby also contribute to binding specificity and affinity. Second, because interacting residues in PDZ domains can undergo large ligand-dependent conformational changes, variations in binding pocket flexibility may contribute to binding specificity. Such a mechanism has been proposed to explain the different binding specificity of the two highly homologous PDZ domains in NHERF1. Finally, genome-wide analysis of PDZ domain binding suggests that PDZ domain selectivity is also achieved by the cellular and subcellular context of the interaction, and this may actually play a more important role than inherent binding specificity.
Regulation of PDZ Binding
PDZ interactions can be dynamically regulated to control the composition and stoichiometry of different multimeric complexes. Phosphorylation of the binding target is the most common mechanism. This is explained by the fact the P-2 serine or threonine in canonical type I PDZ targets can be a substrate for phosphorylation. In these cases, phosphorylation of the residue creates an energetically unfavorable PDZ ligand. For example, phosphorylation of the COOH-terminal site in the Kir 2.3 channel by Protein Kinase A inhibits its interaction with the synaptic PDZ protein, PSD-95, to regulate the channel. Likewise, phosphorylation of the P-2 serine in the β2 adrenergic receptor uncouples the receptor from the NHERF1 PDZ protein, and disrupts receptor recycling in the post-endocytic pathway.
Phosphorylation of sites within PDZ proteins is emerging as an additional mechanism for modulating PDZ binding. Evidence for this was first provided by observations that the interaction of a PDZ protein, NHERF1 (see below), with CFTR is negatively regulated by phosphorylation of a residue in the second PDZ domain. Phosphorylation of sites in or near the first PDZ domain of NHERF1 also disrupt interaction with the Na-phosphate co-transporter, Npt2a. Phosphorylation of sites that are involved in PDZ–PDZ protein oligomerization has also been observed. This is believed to modulate the extent to which some PDZ proteins can form higher order scaffolding complexes.
Finally, switching interactions with different PDZ proteins can differently regulate the activity and localization of target proteins. This occurs when the target has the capacity to bind to several PDZ proteins that have different properties. For example, TIP-1, a protein that consists of a single PDZ domain and lacks other protein–protein interaction modules, binds to certain target proteins to antagonize the scaffolding functions of canonical PDZ proteins.
Polarized Expression of PDZ Proteins in Epithelial Cells
A number of PDZ proteins are preferentially expressed at polarized membrane domains or within critical sorting compartments ( Figure 14.2 ), where they perform retention/sorting operations and organize local signaling complexes at polarized locales. Examples of PDZ proteins that predominately reside at the basolateral membrane of certain intestinal and renal epithelia include syntrophin (see “Dystrophin-Associated Protein Complex,” below), Lin-7 (see “Lin-7/CASK/SAP97,” below), the ErbB interacting protein, ERBIN, and certain members of the membrane associated guanylate kinase family of PDZ proteins, such as CASK, PSD-93, and SAP97 (aka Discs large homolog 1 ). Other PDZ proteins, including the sodium hydrogen exchange regulator factors (see “NHERF, “below), Shank2E, and PSD-95, are chiefly expressed on or near the apical membrane. Some PDZ proteins, such as zonula occludens, PALS1 (Stardust), and PATJ (Disc lost), play important roles in the generation and maintenance of the tight junction. Still others, like CAL, which is primarily located in the Golgi or SNX27, and syntenin, which are found in endosomes, reside in biosynthetic or endocytotic sorting compartments.
A PDZ-binding motif can serve as a polarized sorting or retention signal. One of the first examples evolved from studies with the GABA transporters or GATs ; deletion of the PDZ-binding motif from the apical isoform GAT-3 caused the transporter to localize randomly to both apical and basolateral membranes. Basolateral membrane expression of several membrane proteins has also been found to require a PDZ-binding motif. For instance ERBB receptors, which play crucial roles in morphogenesis and oncogenesis, interact with a basolateral PDZ protein, called ERBIN, and require a PDZ-binding motif for basolateral membrane expression. ERBIN is targeted to the basolateral membrane by its leucine-rich repeat domain. Efficient basolateral membrane expression of a number of transporters that interact with the basolateral PDZ protein Lin-7 also require an intact PDZ-binding site (see below).
MAGUKs, the Archetypal PDZ Scaffolds
Members of the MAGUK (membrane associated guanylate kinase) family of PDZ proteins are the archetypal PDZ scaffolds. MAGUK proteins are equipped to assemble large molecular complexes, having one to three PDZ domains, a SRC homology 3 domain (SH3), and a catalytically inactive guanylate kinase-like (GK) domain. In addition to the PDZ domains, the GK and the SH3 domains function as independent protein–protein interaction modules; GK domains recruit scaffold adaptor molecules called guanylate kinase-associated proteins or GKAPs, while SH3 domains have been shown to coordinate interaction with at least one non-receptor tyrosine kinase. The SH3 and GK domains can also interact with one another, forming a composite SH3–GK structure that acts as an additional intermolecular protein–protein interaction domain with a binding specificity that is distinct from either SH3 or GK domains.
The PSD-95 family, encoded by four genes (PSD-95/SAP90, PSD-93/Chapsyn-110, SAP102, and SAP97), exemplifies MAGUK proteins. Two of these, PSD-93 and SAP97 (see below), are expressed in renal epithelial cells. However, the best characterized member, PSD-95, is largely expressed in excitable tissues, and plays central roles in maintaining and modulating the strength and structure of glutamatergic synapses. Generally, its properties and functions are likely to be applicable to the other MAGUKs, including those expressed in the kidney.
Like many scaffolds, PSD-95 not only contains multiple protein–protein interaction modules, it also assembles into multimers, creating an extended platform for efficient scaffolding. These qualities, combined with palmitoylation-dependent membrane tethering and synaptic localization signals, make PSD-95 ideally designed to cluster ion channels, receptors, trafficking proteins, and signal transduction machinery at the post-synaptic membrane. In doing so, PSD-95 influences trafficking, endocytosis, and activities of target proteins at the synapse. Organizing local signaling complexes is one of the most important clustering functions of PSD-95. For example, the PDZ domains in PSD-95 independently interact with the calcium/calmodulin-activated nitric oxide synthase, nNOS, and NMDA (N-methyl-D-aspartate) receptors to form a ternary complex. The organization is thought to be important for regulated synthesis of nitric oxide. Because NMDA receptors are permeable to calcium, the physical linkage of nNOS with the excitatory receptors is believed to allow nitric oxide production to be efficiently coupled to receptor activation, calcium influx, and local changes in intracellular calcium. Significantly, disruption of NMDAR interaction with PSD-95 dissociates the receptors from downstream neurotoxic signaling, without blocking synaptic activity or calcium influx.
Local signaling complexes that control the production of NOS in the kidney have been proposed. One may involve PSD-93, the predominate MAGUK in renal epithelial cells. Similar to PSD-95, PSD-93 associates with the plasmalemma via palmitoylation-dependent tethering signals, where it recruits and clusters various target proteins, including nNOS. In the kidney, PSD-93 is largely expressed along the basolateral membrane of the thick ascending limb, macula densa cells and the distal nephron. In the macula densa, PSD-93 colocalizes with the pool of nNOS that is associated with intracellular vesicles and the basolateral membrane. It remains to be tested if PSD-93 interaction with nNOS in the macula densa coordinates regulated NO production in the manner that is observed with PSD-95 at the excitatory synapse.
Form and Function of PDZ Protein Families in the kidney
Apical Membrane PDZ Protein Complexes
NHERF
The N a/ H e xchange r egulator f actor PDZ proteins, NHERF, are highly expressed in the kidney and small intestine where they act as molecular scaffolds, associating with a number of transporters, channels, signaling proteins, transcription factors, and receptors to regulate apical membrane transport processes. There is a family of four related NHERF proteins encoded by separate genes ( Figure 14.3 ). Originally known by many names, a unifying nomenclature has been proposed, designating the genes as NHERF-1 (also known as Ezrin Binding Protein-50, EBP-50 ); NHERF-2 (also known as NHE-3 kinase A (E3KARP) ); tyrosine kinase activator-1 (TKA) and sex-determining region of the Y chromosome (SRY-1)-interacting protein ; NHERF-3 (also called PDZK1, Cap70, DiPHOR or NaPi-Cap1 ); and NHERF-4 (also called IKEPP, DIPHOR-2, and NaPi-Cap 2 ). Each member of the NHERF family of proteins is believed to play important roles in the regulation of transport processes within the proximal tubule, as well as other sites along the nephron, acting by three different but not mutually exclusive mechanisms. Present evidence indicates that the NHERFs function to: (1) organize local signaling complexes; (2) control apical membrane trafficking; and (3) couple apical membrane transport proteins with other PDZ-binding targets. In this way, NHERFs modulate transporter activity and/or apical abundance of transporters, channels, and receptors. Importantly, each NHERF isoform appears to have unique regulatory properties that are manifested in cell-specific manners. Studies in NHERF isoform knockout-mice have begun to clarify their different physiologic roles in the renal proximal tubule, small intestine, and other epithelia.
Like other scaffolding proteins, the functions of the NHERFs are made possible by the presence of their multiple protein–protein interaction domains. NHERF-1 and NHERF-2 contain two PDZ domains and a COOH-terminal Ezrin/Radixin/Mosein/Merlin (ERM)-binding domain. The latter coordinates interaction with the ERM family of actin binding and A-kinase anchoring proteins to direct linkage with the actin cytoskeleton and signal transduction machinery. By contrast, NHERF-3 and NHERF-4 contain four PDZ domains, but no ERM domain ( Figure 14.3 ).
NHERF proteins can also interact with one another, forming higher order protein networks. Indeed, NHERF1 and NHERF2 associate as homodimers and heterodimers. Interestingly, oligomerization of NHERF-1, but not NHERF-2, is highly regulated by association with other proteins and by phosphorylation. NHERF3 has been reported to interact with NHERF1 and NHERF2 to form an extensive heteromeric complex. Interaction between NHERF1 and NHERF3 and ezrin is corroborative, providing a mechanism to regulate formation of a ternary scaffolding complex that contributes to the organization microvilli.
NHERF in Epithelial Transport
A growing body of evidence indicates each NHERF isoform has individual and specialized activities in the kidney. It some cases, specific roles of several NHERF proteins may converge and act cooperatively to regulate target proteins. Here we review the state of knowledge about each NHERF isoform.
NHERF1 was originally discovered as a co-factor necessary for cAMP-kinase dependent phosphorylation and inhibition of NHE3, a brush border Na + /H + exchanger. Biochemical studies and work in heterologous expression systems established a likely mechanism whereby NHERF1 organizes a local PKA signaling complex, using its PDZ domains and the ERM-binding domain ( Figure 14.3b ). The second PDZ domain of NHERF1 directly interacts with NHE3, while the ERM-binding domain simultaneously engages ezrin. By acting as an A-kinase anchor protein (AKAP, see below), ezrin recruits the regulatory subunit of PKA II to the NHERF1 complex. Consequently, NHERF1 juxtaposes PKA with NHE3 for efficient phosphorylation of the transporter and inhibition of Na + /H + exchange. Consistent with the model, removal of the ERM-binding domain in NHERF1 disrupts formation of NHERF1–ezrin signal complex and attenuates the inhibitory effect of cAMP on NHE3 activity. EPAC (the exchange protein directly activated by cAMP) also participates in the NHERF1-dependent inhibitory response in the proximal tubule, but it is not presently understood how NHERF couples EPAC to NHE3.
Direct evidence that the NHERF1 signal complex is required for phospho-regulation of NHE3 has been provided by studies in NHERF1 gene knockout mice. In this model, activation of PKA fails to phosphorylate and inhibit NHE3 activity in the proximal tubule. The response appears to be specific to NHERF1 removal, in that other proximal tubule NHERF isoforms are not affected by NHERF1 gene ablation. Moreover, the inhibitory effect of PKA can be completely restored in NHERF1-null proximal tubule cells upon adenoviral-mediated delivery of wild-type NHERF1.
The PKA coupling function of NHERF1 is believed to be widespread, with a body of work indicating that the NHERF1 can act as a nexus of signaling complex assembly for efficient phosphorylation and regulation of a variety of transporters, channels, and receptors (reviewed in ). For example, NHERF1 (as well as NHERF2 ) binds to CFTR through a PDZ interaction to potentiate PKA phosphorylation-dependent CFTR Cl(−) currents in an ezrin-AKAP dependent manner.
Simultaneous PDZ-dependent recruitment of G-protein coupled receptors by NHERF proteins can further focus local signaling around NHE3 and other transport proteins. For instance, studies in heterologous systems reveal that NHERF binds to the β2-adrenergic receptor (BAR2) by means of a PDZ domain-mediated interaction to recruit NHE3 and the receptor into a local signaling complex for efficient receptor-mediated regulation of sodium–hydrogen exchange. Removal of the PDZ interaction motif in the BAR2 disrupts receptor interaction with NHERF1, and markedly reduces β2-adenergic receptor-mediated regulation of NHE3 without altering activation of adenylyl cyclase. Likewise, NHERF1 facilitates the assembly of a complex containing the β2-adenergic receptor, ezrin, PKA, and CFTR at the apical membrane of epithelial cells for compartmentalized and specific signaling of the channel. Other examples have recently been extensively reviewed.
The tandem PDZ domains in NHERF1 also provide a structural framework to link PDZ-binding transport proteins with PDZ-binding signal transduction machinery. Indeed, several different kinases, phospholipase C isoforms, and the receptor for activated C kinase, RACK, have been identified as NHERF1 PDZ-binding targets. Characterization of consensus binding sequences of isolated NHERF-1 PDZ domains by phage-display, affinity selection techniques revealed that the two PDZ domains have different ligand-binding specificities, with distinct preferences for residues at the 0, −1 and −3 positions of type I PDZ ligands. Thus, NHERF1 has a biochemical capacity to tether different PDZ-binding targets together. In addition, because NHERF1 interacts with itself and links with the actin cytoskeleton, formation of an extended network of NHERF1 molecules may join different PDZ-interacting proteins to the same locale. Such a mechanism has been proposed to explain NHERF1-dependent coupling of phospholipase C with the TRP4 channel.
In some cases, the PDZ domains in NHERF1 can support simultaneous interaction of two identical proteins. The best-characterized example is CFTR, which interacts with both PDZ domains in NHERF1, albeit with different binding affinities. In this case, NHERF1 has been reported to induce a high open probability conformation of CFTR by cross-linking the C-terminal tails of a CFTR dimer. Because CFTR binds to the two PDZ domains with different kinetics and affinities, channel gating is profoundly sensitive to alterations in NHERF1 abundance. Moreover, the composition and stoichiometry of NHERF-CFTR interactions can be dynamically regulated. Phosphorylation of NHERF1 has been found to specifically disrupt CFTR interaction with the second PDZ domain, uncoupling the tethered C-terminal tails and inducing a low open-probability conformation. A similar PDZ-dependent cross-linking mechanism has been described with NHERF3.
NHERF1-Dependent Apical Membrane Trafficking
In addition to co-localizing key components of signal transduction pathways, NHERF1 can also regulate cell surface expression and localization of some of its binding targets. It appears to function by controlling trafficking operations in the post-endocytic recycling pathway, as well as by anchoring target proteins on the plasma membrane, likely by interactions with the cytoskeleton.
Regulation of the Na-dependent phosphate transporter, Npt2a, in the proximal tubule provides a salient example. It is well-known that factors which regulate proximal tubule Pi reabsorption and Pi homeostasis do so by altering the density of Npt2a at the apical membrane. NHERF1 plays an important role in this process. Indeed, NHERF1 gene ablation causes diminished expression of Npt2a at the apical membrane and renal phosphate-wasting. In the NHERF1 knockout model, the Npt2a transporter is misrouted into a subapical, intracellular compartment, indicative of a trafficking defect.
Studies in model systems have begun to cast light on the underlying mechanism. Npt2a binds to the first PDZ domain of NHERF1 via a type 1 interaction, requiring the last three amino acids of the co-transporter. These residues are also necessary for efficient apical expression of Npt2a, suggesting that apical targeting and/or anchoring is specified by direct NHERF1 interaction. Apical localization of the co-transporter can be blocked by ectopic expression of truncated NHERF1 proteins, which contain the first PDZ domain and are able to interact with the transporter, but lack the ERM-binding domain. Thus, it is likely that NHERF1 coordinates localization of the co-transporter by tethering Npt2a with the actin cytoskeleton through the ERM-binding domain.
Exciting recent studies reveal that PTH induced internalization and lysosomal degradation of Npt2a in the proximal tubule are coincident with phosphorylation of NHERF1, and disruption of Npt2a/NHERF1 interaction. These observations strongly suggest that Npt2a–NHERF1 interactions are physiologically regulated to control Npt2a apical surface density for maintenance of calcium and phosphate metabolism. Recent live-cell imaging studies indicate NHERF1 regulates apical expression of Npt2a by a brush border retention mechanism. A similar process controls the localization of the PTH receptor. In other proteins, such as CFTR and certain G-protein coupled receptors, NHERF1 maintains surface expression by driving recycling to the cell surface after internalization.
Although NHERF1 effectively anchors NHE3 and Npt2a within the microvilli by directly interacting with the transporters and engaging the underlying microvillar cytoskeleton, the transporters have different fates in the renal proximal tubule when they disassociate from NHERF1. In NHERF1 knockout mice, localization of NHE3 is maintained within the microvilli, but Npt2a is targeted to the lysosome. By contrast, when interaction with NHERF1 become severed by physiological signaling processes (e.g., PTH-dependent), a myosin VI driven translocation process moves NHE3 and Npt2a out of the microvilli. Because NHE3 selectively assimilates with lipid rafts, the translocated NHE3 molecules are effectively excluded from clathrin-coated pits and consequently are retained at the base of the microvilli. By contrast, Npt2a transporters do not partition into rafts, and are free to be internalized once its ties with the microvillar anchor are broken.
Direct phosphorylation of NHERF1 is emerging as an important mechanism for negatively regulating PDZ-dependent binding interactions in the proximal tubule. Two residues, threonine 95 and serine 77, within the first PDZ domain are phosphorylated in response to PTH and dopamine treatment in the renal proximal tubule. This decreases the binding affinity for the Npt2a transporter, and likely contributes to the hormonal suppression of renal phosphate transport. Additional phosphorylation-dependent mechanisms have been reported to control binding at the second PDZ domain in ways that are important for regulating microvilli assembly.
NHERF2
NHERF2 appears to have different functions than NHERF1 in the kidney, even though the two PDZ proteins share a common domain structure and NHERF2 is equally effective as NHERF1 in mediating cAMP inhibition of NHE3 in heterologous systems. Unlike NHERF1, which is exclusively expressed in the human, rat, and mouse proximal tubule, expression of NHERF2 in the proximal tubule is species-specific. Found only in the mouse proximal tubule, NHERF2 predominantly localizes to a subapical, intermicrovillar compartment that is distinct from NHERF1 in the brush border. Importantly, NHERF2 does not support phosphorylation-dependent inhibition of NHE3 or apical localization of Npt2a in the NHERF1 knockout model, indicating that NHERF2 does not share physiologically redundant functions with NHERF1 in the proximal tubule.
The functions of NHERF2 are, in fact, better understood at sites outside the proximal tubule. In the kidney, NHERF2 is predominately expressed in the glomerulus, vas recta, and the collecting duct. Physiologically important PDZ-binding partners have been identified in each of these locales. In the glomerulus, NHERF2 interacts with podocalyxin, possibly functioning to retain podocalyxin at the apical surface of the podocyte and provide a mechanism for linking this important surface sialomucin to the actin cytoskeleton. NHERF2 associates with the TRPC4 channel in the descending vasa recta, where it has been suggested to control Ca 2+ signaling in a similar way that the INAD PDZ protein controls TRP in the Drosophila eye. In the collecting duct, NHERF2 co-localizes and interacts with the ROMK channel. Studies in heterologous expression systems indicate that NHERF2 couples accessory proteins and signal transduction machinery to ROMK for efficient channel regulation and trafficking.
The PDZ-binding specificity of NHERF2 also undoubtedly contributes to its unique functions as compared to NHERF1. While NHERF2 shares many of the same PDZ-binding partners as NHERF1, with nearly 60 having been identified, it also interacts with several proteins that NHERF1 does not react with. These include alpha-actinin-4 ; cGMP kinase I and II ; a putative Cl/HCO 3 exchanger downregulated in adenoma ; podocalyxin ; human Y-linked testis determining gene-binding factor ; serum glucocorticoid stimulated kinase, SGK-1 ; and transcriptional co-activation with PDZ-binding motif, TAZ.
By organizing these unique partners into protein complexes, NHERF2 can affect functions that are distinct from NHERF1. NHERF isoform-specific regulation of NHE3 in heterologous systems provides an excellent illustration. NHERF2 uniquely confers Ca 2+ -dependent inhibition on NHE3 by scaffolding the exchanger to PKCα and alpha-actinin-4. NHERF1 does not support this activity, presumably because it is not capable of interacting with alpha-actinin-4. Likewise, by acting as a unique protein kinase G-anchoring protein, NHERF2 specifically confers cGMP inhibition on NHE3. Finally, activation of NHE3 by dexamethasone requires NHERF2 rather than NHERF1. In this case, the first PDZ domain of NHERF2 uniquely recruits the serum- and glucocorticoid-induced protein kinase, SGK1, into a complex with NHE3 to phosphorylate and enhance exchanger activity. Such a mechanism has been suggested to offer an explanation for glucocorticoid stimulation of sodium absorption in ileum, proximal colon, and renal proximal tubule.
NHERF2-dependent scaffolding of SGK1 may also play an important role in the collecting duct for the regulation of the potassium secretory channel, ROMK. It has been reported that NHERF2 can synergize with SGK1 to augment cell surface expression of ROMK in oocyte expression experiments. Biochemical studies indicate that NHERF2 has the capacity to recruit ROMK and SGK-1 into a ternary complex by preferentially binding to the channel with the first PDZ domain, while simultaneously recruiting the kinase by preferred interaction with the second PDZ domain. Formation of such a complex would allow efficient phosphorylation of a residue that is required for delivery of the channel to the cell surface. Indeed, SGK1 directly phosphorylates serine 44 in ROMK1, creating a forward trafficking signal that overrides an endoplasmic reticulum localization signal. Together the observations suggest a potential molecular mechanism for the regulation of ROMK density by dietary potassium, whereby the NHERF2 scaffold juxtaposes the SGK-1 with ROMK for efficient phosphorylation-dependent trafficking to the apical membrane.
NHERF3
NHERF3 was first discovered as PDZK1, a PDZ domain-containing protein that is upregulated in carcinomas, and abundantly expressed in the proximal tubule brush border. Significantly, the four PDZ domains of NHERF3 support interaction with many proximal tubule apical membrane transporter proteins, including Npt2a, the solute carrier SLC17A1 (NaPi-I), NHE3, the organic cation transporter (OCTN1), chloride-formate exchanger (CFEX), and the urate-anion exchanger (URAT1), as well as a protein kinase A anchoring protein, D-AKAP2. Based on these observations and findings that NHERF1 can interact with NHERF3, it has been suggested that NHERF3 and NHERF1 may form an extended scaffolding network in brush borders of proximal tubular cells for the regulation of transport.
Although the NHERF3-NHERF1 scaffolding network concept is an attractive hypothesis, it should be pointed out that targeted disruption of the NHERF3 gene by homologous recombination does not cause global alterations in the expression or localization of most of its interacting transport proteins in the proximal tubule. Instead, the major effects of NHERF3 gene disruption presently appear to be very specific, confined only to two interacting proteins. A selective reduction in the abundance and functional activity of the chloride-formate exchanger, CFEX, at the proximal tubule brush border is observed in NHERF3-null animals. Physiologic stimulation of the NaPi-IIc isoform by dietary phosphate restriction is also impaired in NHERF3-null animals. A minor role of NHERF3 in Npt2a regulation can be provoked by physiological perturbations; while NHERF3 null animals on a normal or low phosphate diet do not exhibit alterations in Npt2a abundance or function, high dietary phosphate unmasks a modest attenuation of Npt2a levels at the proximal tubule brush border. Differences in affinities of the NaPi-II isoforms for NHERF1(Npt2a) and NHERF3 (Npt2c) have been proposed to account for this behavior.
NHERF4
This member of the NHERF family was originally identified in independent screens for PDZ-binding partners of the Npt2a transporter and the receptor guanylyl cyclase. First dubbed as NaPi Cap-2 or IKEPP (Intestinal and Kidney-Enriched PDZ Protein), it was subsequently reclassified as NHERF4 based on sequence homology modeling. In the proximal tubule where it is abundantly expressed, NHERF4 localizes to a subapical region, like NHERF2, that is distinct from the brush border and NHERF1. Little is known about the function of NHERF4 except that it inhibits heat stable toxin induced cGMP synthesis, by a mechanism suggested to involve PDZ-dependent recruitment of inhibitory factors. Roles in regulating the TRPV5 and TRPV6 calcium channels have also been described. Based on its structural similarities with other NHERF forms and the site of expression in the kidney, it seems likely that NHERF4 also modulates apical membrane transport and cell signaling in the proximal tubule.
Shank (SH3 Domain and Ankyrin Repeating Proteins, A.K.A Proline-Rich Synapse-Associated Protein-1/Cortactin-Binding Protein 1 (ProSAP1/CortBP1))
In addition to the established role as master scaffolds at the postsynaptic density, members of the Shank (SH3 domain and ankyrin repeating proteins) family of proteins play roles as apical membrane-associated scaffolds in epithelial cells. The three known Shank genes (Shank1, Shank2, and Shank3) are expressed in a tissue-specific manner. Shank 1 is almost exclusively expressed in the brain, products of the Shank2 gene are found in brain, kidney, and liver, whereas Shank 3 is most abundantly expressed in the heart. The prototypical Shank, Shank1, is a relatively large protein (>200 KDa) containing multiple ankyrin repeats, a SH3 domain, a PDZ domain, and a long proline repeat domain. A self-oligomerization domain, called a sterile alpha motif (SAM), assembles the scaffolds into head-to-tail helical sheets, forming an extensive Shank network for protein complex assembly. In neurons, Shank forms a polymeric structure with another protein, Homer, to serve as a platform for assembling postsynaptic density proteins. Multiple splice variants of each gene have been identified that contain different combinations of protein–protein interaction domains ( Figure 14.4 ). For example, Shank2E, a form that is predominantly expressed in epithelial cells, contains ankyrin repeats whereas the Shank2 splice form found in the brain does not.
Extensive studies in the brain provide models for Shank function in epithelial cells. It is well-established that Shank family members localize to the postsynaptic density of excitatory synapses, where they act as master scaffolds along with Homer and PSD-95. Here, the Shanks physically couple the two major receptor complexes, N-methyl-d-aspartate receptors (NMDAR) and metabotropic glutamate receptors (mGluR), and recruit associated signaling proteins. They do so by concurrently engaging two different adaptors through distinct protein–protein interaction domains. The Shank PDZ domain associates with the GUK-associated protein, GKAP which, in turn, interacts with the PSD-95 complex, containing NMDA receptors. At the same time, the proline rich domain of Shank interacts with another adaptor protein, Homer, to link with metabotropic glutamate receptors. Proline rich domains often serve as binding sites for SH3 (SrC homology), WW (conserved two-tryptophan domain), and EVH1 domains (enabled/vasodilator-stimulated phosphoprotein homology 1). A single EVH1 domain in Homer directly interacts with a PPXXF motif in the Shank proline-rich domain, as well as with similar proline motifs in group 1 mGluR and other proteins, such as the IP3 receptor. Because Homer proteins self-associate in a head-to-tail fashion, two EVH1 domains per dimer are available to bridge Shank with group 1 mGluRs. Consequently, Shank cross-links Homer and PSD-95 complexes in the PSD, presumably to couple signaling transduction pathways emanating from NMDAR and mGluR. Disruption in this synaptic scaffolding mechanism may be responsible for human disease, as mutations in SPAK2 have been associated with autism and mental retardation.
More recently, Shank2 forms have been implicated in the modulation of apical membrane transport processes. In the kidney, Shank2E is concentrated at the apical membrane of proximal tubule cells were it interacts with NHE3 and Npt2a, similar to NHERF1. Present evidence indicates that Shank2 and NHERF may control the activity of these transport proteins in divergent manners. Studies with NHE3 in heterologous expression systems, for example, revealed that Shank2 positively regulates NHE3 membrane expression and blunts the cAMP-dependent inhibition of NHE3, in part by antagonizing the action of NHERF1. Likewise, in pancreatic duct cells, Shank2E associates with CFTR at the apical membrane and inhibits Cl channel activity, contrasting the positive effects of NHERF1 and NHERF2 on CFTR (see above). Shank2 also positively regulates NHE3 by recruiting BetaPix, a guanine nucleotide exchange factor for the Rho-GTPase. Because retention and targeting of NHE3 in the apical microvilli depends on the sustained activity of Rho-GTPases, the interaction between NHE3 and the Shank2-BetaPix complex may allow NHE3 trafficking to be linked to the maintenance of the microvilliar actin cytoskeleton.
Shank2E appears to regulate Npt2a in a different manner than NHERF1. In the proximal tubule, increased extracellular Pi triggers internalization and degradation of Shank2E and Npt2a in parallel, but has no effect on NHERF1 localization or abundance. Combined with observations that regulated endocytosis of Npt2a is associated with disruption of Npt2a/NHERF1 interaction at the brush border, one might speculate that internalization of Npt2a involves a NHERF1-to-Shank2E interaction switch. Importantly, Shank redistributes with Npt2a during regulated endocytosis, and interacts with dynamin II, a GTPase that is critical for endocytic vesicle formation, via proline-rich domain interaction. Thus, Shank2E is especially poised to facilitate Npt2a endocytosis and/or lysosomal trafficking, in contrast to the apparent membrane-retention and/or recycling function of NHERF1. The molecular mechanisms underlying the function of Shank2E in the proximal tubule remain to be firmly established, however. It will be interesting to learn if the activities of Shank2E in the kidney depend on scaffold adaptors, such as Homer and GKAP, as has been shown in excitatory synapses.
Basolateral Membrane PDZ Protein Complexes
The Lin-7/CASK/PSD-97 System
Lin-7 and CASK (Lin-2) are components of an evolutionarily conserved basolateral membrane scaffolding complex, important for polarized targeting and controlling cell surface density of their PDZ-binding partners. They were discovered along with another PDZ protein, Lin-10 (Lin, from abnormal cell lineage), in a genetic screen for components of the LET-23 receptor tyrosine kinase signaling pathway in C. Elegans vulva progenitor cells (VPC). These molecules form a tripartite protein complex in VPC that interacts with a receptor tyrosine kinase, LET-23, to coordinate receptor expression on the basolateral membrane. Importantly, null mutations in Lin-7, Lin-2 or Lin-10 cause the Let-23 receptor to become mislocalized to the apical membrane, and consequently disrupt LET-23 signaling and VPC development.
Orthologs of the C. elegans PDZ protein complex have been identified in mammalian tissues (Lin-7=mLin7/Veli/MALS; Lin-2=CASK; Lin-10=Mint-1/X11). In the mammalian kidney, a partially conserved complex, consisting of mLin-7 and CASK but not Lin-10, localizes to the basolateral membrane where it coordinates polarized expression of mLin-7-binding partners ( Figure 14.5 ). It has been implicated in basolateral expression of the epithelial GABA transporter, BGT-1, the strong inward-rectifying potassium channels, Kir 2.X, the EGF-like receptor, ErbB-2/Her2, and the insulin receptor substrate, p53. Present evidence suggests that the mLin-7/CASK complex may offer a general mechanism for polarized expression of basolateral membrane proteins containing Type I PDZ-binding motifs.
Lin-7 acts as the upstream scaffolding molecule. It binds directly to target molecules through a Type I PDZ interaction while simultaneously engaging CASK via another protein–protein interaction module, called a L27 domain (from L in- 2 , Lin- 7 ). L27 domains, which present a number of related PDZ proteins (see PALS, below), are helical bundle structures that mediate heterotypic assembly, important for polymerization of different scaffolds. In fact, basolateral membrane localization of Lin-7 and its PDZ-binding partners is afforded by the L27 domain and CASK interaction.
CASK associates with the basolateral membrane through a web of interactions to function as the master basolateral membrane attachment factor. As a member of the MAGUK family (see above), CASK contains multiple protein–protein interactions sites, allowing it to simultaneously bind to Lin-7, extracellular matrix receptors, adhesion molecules, the actin cytoskeleton 4.1-binding proteins, and another MAGUK protein, SAP-97. The mLin-7 and SAP-97 L27 domains separately assemble with two L27 domains of CASK, possibly as a dimer of L27 heterodimers, to form a mLin-7/SAP97/CASK complex. By linking extracellular matrix receptors and the cytoskeleton, the Lin-7/CASK/SAP97 complex has the capacity to act as a stable anchor to retain Lin-7 interacting proteins on the basolateral membrane.
The mammalian counterpart of Lin-10 is actually encoded by a family of proteins called the Mints or X11s. Although all three members of the Mint family share C-terminal PDZ and PTB domains, only Mint-1 contains a CASK interaction domain. In neuronal tissues and the heart, which express the complete Lin-7/CASK/Mint-1 complex, Mint-1 has been suggested to provide an additional membrane trafficking function. Mint-1 interacts with microtubule motors, and has been reported to transport N-methyl-D-aspartate (NMDA)-type receptor vesicles along microtubules. In addition, Mint-1 interacts with Munc-18 docking machinery. Importantly, mammalian epithelial cells do not express Mint-1. In its absence, the Lin-2/CASK system loses the obvious link to microtubule-mediated trafficking and fusion, suggesting that Lin-7/CASK plays a major role in retention rather than directed-delivery in renal epithelia.
Consistent with this notion, present evidence indicates that Lin-7/CASK primarily functions to retain target proteins at the basolateral membrane of mammalian epithelia. For example, Perego et al. found that removing the PDZ ligand in BGT-1 disrupted Lin-7 association in MCDK cells and dramatically increased the internalization of the transporter from the plasmalemma. The retention function of Lin-7 depends on its L27 domain, which directs interaction with a cognate L27 domain in CASK. In this way, Lin-7 acts as a PDZ-to-L27 adapter, mediating indirect association of its PDZ-binding target proteins with the larger basolateral membrane scaffold, CASK.
Although Lin-7 primarily operates as a component of a basolateral membrane retention machine in mammalian epithelial cells, it should be pointed out that disruption of Lin-7 interactions can produce a wide range of mis-localization phenotypes, depending on the Lin-7-binding partner and the types of sorting signals embedded within them. For instance, mutant BGT transporters, lacking their PDZ-binding motif, are predominately localized on the basolateral membrane. In this case, BGT-1 transporters are presumably directed to the basolateral membrane by non-PDZ-dependent sorting signals. On the other hand, mutant Kir2.3 channels, lacking the PDZ-binding motif, are largely directed to an endosomal compartment, rather than the basolateral membrane, consistent with strong endosomal targeting signals. An apical-missorting phenotype is produced by removing the PDZ-binding site from a chimeric LET-23/nerve growth factor receptor protein. In this case, Lin-7 interaction may stabilize the receptor on the basolateral membrane or limit post-endocytic trafficking in such a way that it prevents transcytosis to the apical membrane.
At present, it is not known if the Lin-7/CASK/SAP97 scaffold also participates in signal complex localization and organization at the basal membrane in the way that has been described for the NHERF proteins at the apical membrane. SAP97 has been shown to recruit AKAP proteins to related PDZ protein complexes in neurons, and the Drosophila CASK ortholog protein, Camguk, functionally modulates Ether-a-go-go potassium channels by a phosphorylation-dependent mechanism. Interestingly, CASK contains an unusual CAM-kinase like domain, which can phosophorylate interacting proteins at the synapse. However, it remains to be established if similar mechanisms are in place at the basolateral membrane of renal epithelial cells, which express the Lin-7/CASK/SAP97 complex.
Lin 7 Isoforms (Veli/MALS) and PALS (Partners of Lin7) in the Kidney
Three different Lin-7 isoforms, encoded by separate genes, have been identified. In mammalian systems, these are often called Veli/MALS 1,2,3 ( Ve rtebrate Li n-7 or Ma mmalian L in S even). Each of the MALS/Veli isoforms are expressed in the kidney, but each are differentially localized along the nephron. MALS/Veli 1 is predominately expressed in the glomerulus, thick ascending limb of Henle’s loop (TAL), and the distal convoluted tubule (DCT). MALS/Veli 2 is exclusively expressed in the vasa recta. MALS/Veli 3 is largely located in the proximal tubule, DCT, and collecting duct. The subcellular localization of MALS/Veli proteins can vary, depending on the isoform and the cell type. In contrast to the predominate basolateral location of MALS/Veli 1 in the TAL and DCT and MALS/Veli 3 in the DCT, MALS/Veli 1 is found diffusely throughout the cytosol of intercalated cells. In the collecting duct, MALS/Veli 3 is chiefly located on the basal membrane. Collectively, these results suggest that different MALS/Veli isoforms may carry out cell type-specific functions. MALS/Veli 1 and 2 isoforms in the thick ascending limb and distal segments appear to have the most significant capacity for a basolateral membrane targeting mechanism. MAL3/Veli 3 has been implicated in generation of polarity in the proximal tubule.
The disparate subcellular localization patterns of MALS/Veli isoforms are likely to arise from at least two different factors. First, differences in the primary structures provide reason to suspect that the MALS/Velis may have specific binding preferences. In contrast to the nearly identical PDZ domains amongst the MALS/Veli isoforms, the extreme NH 2 – and COOH-termini are highly divergent. Most importantly, the region believed to direct MALS/Veli subcellular localization, the L27 interaction module, exhibits only 57% amino acid identity between isoforms, raising the possibility that different isoforms preferentially bind to different L27 domain proteins. Second, cell-specific expression of Lin-7-binding partners may account for differences in isoform localization in different nephron segments and cell-types. A group of CASK-like MAUGK proteins, called PALS (Partners of Lin-7), that contain L27 hetero-oligomerization domains, have been identified as potential partners of Lin-7. The different PALS might substitute for CASK under certain circumstances, forming MALS/Veli complexes with different subcellular locations and disparate functions. For instance, PALS1 targets Lin-7 to the tight junction, in contrast to the basolateral membrane location of the CASK/Lin-7 complex.
Dystrophin-Associated Protein Complex
The dystrophin-associated protein complex (DPC) is a transmembrane scaffolding machine that is expressed in a variety of tissues. Extensive studies in skeletal muscle revealed the DPC serves structural and signaling functions. In epithelia, the DPC localizes to the basolateral membrane ( Figure 14.6 ), where it is organized in a manner similar to the skeletal muscle complex, functioning to compartmentalize and tether signaling and transport proteins.
Related posts:
Solute Transport, Energy Consumption, and Production in the Kidney
Mechanisms of Ion Transport across Cell Membranes
Molecular and Cellular Mechanisms of Kidney Development
Physiopathology of Potassium Deficiency
Clinical Disturbances of Phosphate Homeostasis
Pathogenesis and Treatment of Nephrolithiasis
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
