Holding open the door reveals a new view of polycystin channel function

Abstract

The functions of polycystin 1 and polycystin 2 (PC1 and PC2) have been surprisingly difficult to establish. PC1 and PC2 are encoded by the Pkd1 and Pkd2 genes that are implicated in autosomal dominant polycystic kidney disease (ADPKD). ADPKD is the most common potentially lethal genetic disorder, affecting ~1 in 1,000 people. Over the course of decades, ADPKD patients’ kidneys acquire numerous fluid‐filled cysts whose expansion compresses the surrounding parenchyma, leading to end‐stage renal disease in ~50% of afflicted individuals [1]. Identification of the genes encoding the PC proteins 20 years ago led to the hypothesis that they form an ion channel, since the sequence of PC2 marks it as a member of the TRP family of cation channels. In the ensuing 2 decades, tremendous effort has been devoted to determining whether this is indeed true and, if so, what characteristics that channel might manifest. A recent paper by Wang et al in this issue of EMBO Reports [2] demonstrates that assembly with PC1 changes the properties of the polycystin channel in ways that may help explain the complex behaviors that have been attributed to it.

The complexity inherent in unraveling the functions of the PC proteins stems from 2 sources. The first of these relates to the confusing and often contradictory behaviors of the PC proteins themselves. They appear to occupy a multiplicity of subcellular locations individually and together, to participate in a variety of interactions with themselves and with each other, and to contribute to the formation of ion channels whose fundamental attributes have been the subject of much debate. The second derives from the difficulties inherent in proving that a protein is indeed part of a novel ion channel and in characterizing its biophysical features. The paper in this issue by Wang et al, which demonstrates quite elegantly that the complex formed by PC1 and PC2 can function as a calcium‐permeable ion channel, does a wonderful job of addressing both of these sets of imposing challenges 2.

Members of the Trp family of ion channels generally form homotetramers, all the subunits of which contribute 2 of their 6 helices to the creation of the channel pore. PC2, when expressed in the absence of PC1, forms just such a homotetramer 3, 4. The interactions among the monomers in the PC2 homotetramer are remarkably cozy, involving domain swaps in which each monomer contributes its 2 pore‐lining helices to its neighbor to form an interlocking ring. PC1 has 11 transmembrane spans, the final 6 of which exhibit homology to the transmembrane domains of PC2. Previous results strongly suggested that PC1 and PC2 can form a heterotetramer that incorporates 3 PC2 monomers and 1 PC1 protein 5. The recent high‐resolution structure of the complex between PC1 and PC2 reveals that this stoichiometry is correct and that it exhibits an arrangement similar to that of the PC2 homotetramer 6. In this case, PC1 contributes its final 2 pore‐forming helices (10 and 11) to its PC2 neighbor and accepts in turn helices 5 and 6 from the PC2 protein on its other side (Fig 1). Interestingly, this structure indicated the presence of a pore‐blocking positively charged kink in the final 2 helices of PC1, suggesting that the conformation that was resolved does not conduct cations.

Figure 1. Schematic diagrams of the PC2 homotetramer and the PC1 and PC2 heterotetramer.

A schematic top‐down view of the structures of the domain‐swapped homotetramer formed by PC2 alone (A) and the similarly organized heterotetramer formed by 3 PC2 proteins and 1 PC1 (B). The data in Wang et al suggest that the PC2 homotetrameric channel conducts K⁺, whereas the PC1/PC2 heterotetrameric channel conducts K⁺ and Ca²⁺.

PC1 and PC2 exhibit overlapping non‐identical distributions when examined in native tissues and following heterologous expression in cultured cells. Perhaps the most interesting site of overlap is the primary cilium. It has been suggested that the ciliary PC proteins collaborate to form a calcium‐permeable channel that is activated by flow‐induced mechanical bending or by chemical or hormonal stimuli. Because of the primary cilium's minute size, it has been difficult to define directly the cilium's electrophysiological properties, although several recent tour de force efforts have succeeded in exploring these parameters. The results of these analyses are somewhat complicated. To summarize imperfectly and overly simplistically this very large body of data, it would appear that cilia in renal epithelial cells contain a channel that is permeable to monovalent cations and calcium and whose existence is dependent upon the expression of PC2 7. Furthermore, under circumstances in which PC2 is expressed by itself, in vivo or in vitro, the observed cation channel conducts monovalent cations and exhibits extremely low calcium permeability 8. These results suggest the intriguing possibility that PC2 can participate in the formation of at least 2 different channels with distinct subunit compositions and ion permeability properties.

It seems like a trivially simple tautology to state that one cannot study the ionic conductivity of an ion channel unless ions can flow through it. Embedded within this deceptively simple assertion, however, is one of the principle challenges associated with the characterization of a novel ion channel. To study what flows through a channel when it is open, one needs to know how to open it. While several studies have identified conditions and ligands that may activate the PC channel, no clear consensus has emerged regarding a specific and biologically relevant channel activator. Thus, it has been challenging to prove that a channel activity detected in native tissue or in a heterologous expression system is, in fact, attributable to a pore formed by PC2 alone or in association with PC1. Substituting a single amino acid residue in the sequence of PC2‐F604P results in the production of a channel that is constitutively open when expressed by itself in Xenopus oocytes 9. This channel is far more permeable to potassium than it is to calcium, which actually blocks the channel when it is present at high concentrations in the extracellular medium. Wang et al now show that co‐expression of PC1 with PC2‐F604P dramatically reduces the current that was detected in comparison with that measured when PC2‐F604P is expressed by itself 2. They find, however, that substitution of 2 residues near the putative gating domain of PC2, to produce PC2‐L677A/N681A, results in a protein that forms a channel when co‐expressed with PC1. Furthermore, the channel produced through this co‐expression manifests a small but significant permeability to calcium. PC1 co‐assembles with PC2‐L677A/N681A, and substitution of a putative pore‐lining residue of PC1 (R4090W) eliminates the channel activity of the co‐expressed proteins. Taken together, these findings provide strong support for the conclusion that heterotetramers formed between PC1 and PC2 can act as calcium‐permeable channels, while PC2 homotetramers form channels that are primarily permeable to monovalent cations (Fig 1).

This important discovery may help to resolve the conflicts that have accumulated in the polycystin channel literature. Like any worthwhile novel observation, however, it also inspires a cascade of questions. Perhaps paramount among these is whether PC2 participates in the formation of these 2 biochemically and electrophysiologically distinct types of channels in vivo. The concept that PC2 may function differently in the presence or absence of PC1 makes perfect sense in light of the fact that PC1 is expressed at extremely low levels in renal epithelial cells following the completion of kidney development, whereas PC2 expression persists throughout adulthood 10. It is logical to assume that this large and lonely population of PC2 must serve a physiological purpose in the absence of stoichiometric quantities of PC1. If, in fact, PC2 can participate in the formation of both homotetrameric and heterotetrameric channels, it is natural to wonder whether its decision to participate in the formation of one or the other kind of channel is regulated by signaling mechanisms or environmental stimuli. It is also logical to assume that these 2 different types of channels serve distinct biological functions. It will be critical to elucidate when these different types of channels form during development, where they are localized within cells and along the nephron, and whether their relative expression is altered in response to various cues. In addition, it will be imperative to determine how these 2 channel types are activated in their native setting and to assess whether their gates are controlled by the same or different factors. From a biophysical perspective, it will be interesting to define the single‐channel behaviors of these 2 different PC channel types and to assess whether the presence or absence of PC1 influences their kinetic as well as their selectivity properties. Thus, the observations reported in Wang et al have the potential to open an exciting new chapter in ADPKD research that may help us to understand better what roles these proteins are supposed to play as well as why cystic disease is the phenotype that develops in their absence.

EMBO Reports (2019) 20: e49156