The Birth of a Channel

Neuron, Vol. 40, 265–276, October 9, 2003, Copyright 2003 by Cell Press

The Birth of a Channel

Carol Deutsch* Department of Physiology University of Pennsylvania Philadelphia, Pennsylvania 19104

An ion channel protein begins life as a nascent peptide inside a ribosome, moves to the endoplasmic reticulum where it becomes integrated into the lipid bilayer, and ultimately forms a functional unit that conducts ions in a well-regulated fashion. Here, I discuss the nascent peptide and its tasks as it wends its way through ribosomal tunnels and exit ports, through translocons, and into the bilayer. We are just beginning to explore the sequence of these events, mechanisms of ion channel structure formation, when biogenic decisions are made, and by which participants. These decisions include when to exit the endoplasmic reticulum and with whom to associate. Such issues govern the expression of ion channels at the cell surface and thus the electrical activity of a cell. The defining feature of a neuron is its excitability, waves of electrical signals generated by the activity of ion channels. Functionally, these complex proteins have been well understood for many decades, especially since the advent of techniques for tracking ion movement (voltage clamp) and protein movement (fluorescence and cysteine scanning accessibility). More recently, we have witnessed the proliferation of crystallographic and cryoEM structures for a variety of ion channels or their subdomains, as well as biochemical evidence that channels are macromolecular complexes of different types of proteins. While the architecture has explained many mechanisms of gating and permeation, it leaves unanswered how these complex structures arise. Each pore-forming subunit of a channel begins life as independently translated monomers emerging from individual ribosomes. It is the molecular journey from the earliest points in mRNA translation, where the nascent chain resides within the ribosome, to higher-order assembly and trafficking that is the subject of this review. Retirement, death, and disposal are (as in our lives) deferred. In the Beginning The birth of a channel begins when newly exported mRNA associates with cytosolic ribosomes and tRNA and undergoes translation. This is a multiple birth. A polysome (a string of many ribosomes on a single mRNA molecule) forms and is engaged in the synthesis of multiple copies of the channel protein. The ribosome is composed of two subunits, made of both rRNA and protein. The small subunit decodes mRNA; the large subunit effects protein synthesis via the peptidyl transferase center (PTC) located in a cavity at the interface between the two subunits (Figure 1A). The PTC is composed of RNA, which supposedly functions as a ribozyme to *Correspondence: [email protected]

Review

catalyze peptide bond formation (for structures, see Ban et al., 2000; Beckmann et al., 2001; Gabashvili et al., 2001; Harms et al., 2001; Menetret et al., 2000; Nissen et al., 2000). There is indeed a birth canal, the ribosome tunnel, which begins at the PTC and runs through the large subunit to exit ports some distance away. The nascent channel peptide travels through this birth canal as the peptide is elongated. (“Nascent” is used to denote peptides still bound as peptidyl-tRNA to the PTC of the large ribosomal subunit.) The dimensions of the tunnel, 100 A˚ in length and 10–20 A˚ in diameter, are wide enough to accommodate secondary, e.g., an ␣ helix, which has a 10–20 A˚ diameter, but perhaps not tertiary conformations of the channel protein. Thus, proteins likely acquire their tertiary structures after emerging from the ribosome, a feat that requires at least 30 to 70 amino acids in length for the tethered peptide to traverse the 100 A˚ long tunnel. Proteins still tethered to the PTC can fold and even acquire enzymatic activity shortly after exiting the ribosomal tunnel (Kowarik et al., 2002; Kudlicki et al., 1995). Similar studies are just beginning to be available for nascent K⫹ channels. For example, for Kv1.3, a voltage-gated potassium channel (Kv), the N-terminal domain of the nascent peptide can acquire tertiary structure after exiting the ribosome (Kosolapov and Deutsch, 2003), but not before (A. Kosolapov and C.D., unpublished data). We do not yet know the specific constraints on channel folding, precisely how much folding occurs at the stage of emergence, nor the full cast of enablers. However, ribosomes, specifically 23S rRNA (in the large subunit of prokaryotes), do have chaperone-like activity (Das et al., 1996; Kudlicki et al., 1997). Inside the ribosome, what environment does the protein experience? The question is particularly interesting given that transmembrane segments of an ion channel can be dramatically different. For example, ␣-helical segment S4, the main voltage sensor in a Kv channel, contains several positively charged residues, distributed every third residue around the ␣ helix (Figure 2A). Does it traverse the same path as a strongly hydrophobic transmembrane segment such as S6 in the same Kv protein? Protein as Barcode Once a nascent chain begins to elongate, it can survey its environment, itself, or be the subject of surveillance, thus the twin birth of peptide and signaling potential. The nascent peptide is now a barcode. Amino acid sequence motifs are decoded in the exit tunnel and influence elongation, pausing, and termination steps (Lovett and Rogers, 1996; Morris and Geballe, 2000; Kramer et al., 1999; Tenson and Ehrenberg, 2002), and even the choice of routes to travel and emerge (Choi and Brimacombe, 1998; Clark and King, 2001; Komar et al., 1997). Moreover, a constricted part of the exit tunnel, composed of both protein and RNA, reads the nascent peptide and acts as a discriminating gate to regulate rates of translation (Berisio et al., 2003; Nakatogawa and Ito, 2002). Even more intriguingly, nascent peptides may not traverse the ribosome in a linear fashion but can circle back

Neuron 266

Figure 1. The Ribosome/Nascent Peptide/Translocon Complex (A) The large subunit (L, green) and small subunit (S, purple) are in the cytosol and are bound to the translocon (light blue) in the ER membrane. The tRNA (indicated in black) resides in the peptidyl transferase center (PTC). Emerging from it is the nascent peptide (red), which travels through the ribosome tunnel to the exit site. A chaperone protein (BIP, dark blue) is shown in the ER lumen, in close association with the translocon. mRNA is omitted for clarity. Translocation across the ER membrane is shown in the left image. Note that entry into the lumen is denoted by a clear pathway; BIP (dark blue) is swung out of the way. Integration into the bilayer is shown in the right image. Note that BIP seals off entry into the lumen at some stages during integration. (B) Two models are shown for emergence of cytosolic loops: the ribosome-translocon seal is partially broken (left, Haigh and Johnson, 2002) or a 15 A˚ gap between the ribosome and the translocon is bridged by connectors that form windows through which the cytosolic loop of the nascent peptide exits into the cytosol (right, Menetret et al., 2000).

through the region of the PTC (Choi and Brimacombe, 1998), thus allowing nascent peptide to scan its own new domains as they are made. (NB: Prokaryote and eukaryote ribosomes differ somewhat in structure and regulatory proteins. Thus, some of the above characteristics may apply to one but not the other ribosome.) Upon emergence of a specific amino acid sequence from the ribosome, the entire nascent peptide/ribosome complex is targeted to the ER membrane where synthesis continues. Typically, targeting of membrane proteins involves a signal sequence in the N terminus that binds a signal recognition particle (SRP), which subsequently binds to a receptor on the ER membrane. Since hydrophobicity and helicity of the nascent peptide are critical for its interaction with SRP, and subsequently the ER, acquisition of secondary structure in the ribosome tunnel may be important for efficient ion channel biogenesis. The acetylcholine receptor channel has a cleavable signal sequence (Anderson et al., 1982), whereas Kv channels do not. Rather, Kv channels use transmembrane domains for targeting. For Kv1.3, several individual transmembrane segments are able to target to the ER and translocate across the membrane, while others

also stop translocation across the membrane (Tu et al., 2000). The second transmembrane segment, S2, functions as the signal sequence. Similarly, in the CFTR Cl channel, the second transmembrane segment serves to target the channel to the ER (Lu et al., 1998). Consequently, the first several hundred amino acids are available for a short time in the cytosol and may fold and/or associate with other proteins or factors. However, there is evidence that ER-bound ribosomes can initiate protein synthesis de novo (Potter and Nicchitta, 2000). If so, could insertion of the nascent peptide start with S1? Topologically, the inside of the ER is equivalent to the extracellular space and contains a unique group of proteins and factors that modify the nascent channel peptides, as well as regulate the intralumenal milieu (Figure 2B). Upon arrival at the ER, the ribosome/ nascent peptide complex binds to the translocon, a multiprotein complex that forms an aqueous pore through which the nascent channel peptide enters the ER lumen (Johnson and van Waes, 1999; Figure 1). The ribosome exit tunnel aligns with the aqueous pore of the translocon, which spans the bilayer and only opens to the lumen when a translocating peptide reaches a certain length (Crowley et al., 1994). Two orthogonal events are governed by the translocon: translocation across the bilayer and integration into the bilayer (Figure 1A). The amino acid sequence of the nascent protein, read by the bound ribosome and chaperone proteins, appears to govern these choices (Johnson and van Waes, 1999). For example, a transmembrane segment can cause a pause in translocation. The segment is read as a barcode in the ribosome, and a signal-transducing mechanism communicates with distant gates in the translocon to ensure proper entry (and orientation) into the bilayer. Signals to open and close the translocon pore may be given by sequence motifs that are only four amino acids away from the PTC (Liao et al., 1997). Such signaling is likely to be a fundamental feature of ion channel biogenesis, since ion channels are polytopic membrane proteins. Translocation defines the sidedness (transmembrane topology) of the channel protein and, thus, the location of folding events for cytoplasmic domains and for extracellular/lumenal domains. Topological fate is determined in the translocon. This entails a multitude of intriguing issues. For polytopic ion channels, how do consecutive transmembrane segments orient in opposite directions? Transmembrane segments may insert in pairwise fashion, or consecutive transmembrane segments may invert 180⬚ upon entry or residence in the translocon. How and when is topology decided? Are proteins the only topological determinant, or can lipids help orient the nascent peptide? Topological fate is encoded by the amino acid sequences in the transmembrane and its flanking regions. This encoding has been characterized for a few ion channels: CFTR, Kv1.3, KAT1 (Lu et al., 1998; Sato et al., 2002; Tu et al., 2000). In a channel polypeptide, it is clear that transmembrane segments are topogenically nonequivalent and that multiple topogenic determinants can cooperate in biogenic units to facilitate integration into the bilayer (Sato et al., 2002; Tu et al., 2000) (Figure 2). Deciphering the details of this code and the code readers/executors is our next frontier.

Review 267

Figure 2. Kv Membrane Topology (A) Cartoon of a Kv monomer. The six transmembrane segments (S1-S6) are shown in yellow (one topological configuration, but see Jiang et al., 2003a, 2003b). The N and C termini are in the cytosol. A T1 recognition domain is indicated in green, and a glycosylation site between S1 and S2 is indicated by the ball and stick. Brackets indicate putative biogenic units. Charges (⫺ and ⫹) are indicated for some of the residues in the transmembrane domains of the voltage sensor (S1-S4). The pore region includes S5-P-S6, where P is the intervening loop between S5 and S6. S1-S4 is the voltage sensor domain. (B) Opposite topologies for a Kv monomer in the plasma membrane and the ER membrane are shown. In both cases, the N and C termini are in the cytosol.

When are these topogenic decisions made? One possibility is that a transmembrane segment establishes topology before translation of the subsequent transmembrane segment. Alternatively, as is the case for Kv1.3 (and CFTR), the first transmembrane segment achieves its topology only after the second transmembrane segment is synthesized (Lu et al., 1998; Tu et al., 2000). Interestingly, these topogenic decisions do not have to be irrevocable (Lu et al., 2000). Finally, it is intriguing to consider that protein-lipid interactions may contribute to topological decisions. Because a transmembrane segment is adjacent to both phospholipids and translocon proteins in the translocon (McCormick et al., 2003), charged lipid head groups may actively serve to effect a particular topological or folding event at specific biogenic stages (Wang et al., 2002b). Eventually, the transmembrane segment of a nascent channel peptide exits the translocon and integrates into the hydrophobic interior of the bilayer. A transmembrane segment moves sequentially from the center of the translocon to the periphery, making specific contacts with constitutive translocon proteins along the way (Do et al., 1996; McCormick et al., 2003). Do all transmembrane segments traverse the same route? In one protein, opsin, the relative position of each transmembrane segment in the polypeptide determines which translocon protein associates with it during membrane integration (Meacock et al., 2002). But, this leaves open the question of when, and how many, transmembrane segments leave the translocon at a time. They could leave one transmembrane segment at a time, sequentially, or in

some coordinated fashion (Hegde and Lingappa, 1997; Figure 3). For some proteins, a sequential one-at-a-time “spooling” mechanism makes sense, but for others, it may not. In Kv channels, transmembrane segments S2, S3, and S4 contain negatively and positively charged residues, yet these transmembrane segments are integrated into the hydrophobic bilayer (Tu et al., 2000). How is this possible? Interactions between transmembrane regions may actually be required for them to be stable in the membrane bilayer (e.g., neutralizing charges, satisfying H bonds). In these cases, it makes sense that multispanning transmembrane regions (biogenic units) should be formed and even reoriented within the translocon before integration into the bilayer. An expandable translocon (20–60 A˚; Hamman et al., 1997) can accommodate multiple boarders (Borel and Simon, 1996; Do et al., 1996; Thrift et al., 1991). Not only might the translocon widen during translocation, it may also contract during integration of the nascent peptide into the bilayer (Haigh and Johnson, 2002). For some non-channel proteins, transmembrane segments integrate into the bilayer soon after entering the translocon (Mothes et al., 1997), and hydrophobic signal-anchor transmembrane sequences can contact phospholipid while the rest of the polypeptide is still being translated (Martoglio et al., 1995; McCormick et al., 2003). In Kv1.3 and in KAT1, multiple transmembrane segments interact with each other, resulting in translocation and integration efficiencies for S2-S3-S4 and S5-P-S6 (“P” indicates “pore loop” between S5 and S6) peptide fragments that are markedly

Neuron 268

Figure 3. Integration Models (Top) The “spooling” model depicts sequential integration of each transmembrane segment (heavy black line). (Bottom) In the biogenic unit model, several transmembrane segments are housed within an expandable translocon (light blue) and are later released into the bilayer while subsequent transmembrane segments continue to be made, oriented, and associated with the translocon. In both models, the lumenal gate (BIP, not shown) and ribosome-membrane junction open and close, modulated by sequences in the nascent peptide, while maintaining the permeability barrier of the membrane. Modified after Hegde and Lingappa (Hegde and Lingappa, 1997).

different from the corresponding single transmembrane fragments (Sato et al., 2002; Tu et al., 2000) (see Figure 2). This suggests that S2-S3-S4 and S5-P-S6 are modular biogenic units that form, fold, and orient within the translocon. Indeed, the core of Kv channels is made of two modular functional units, the voltage sensor (S1S4) and the pore (S5-P-S6), and crystal structures of each of these units have recently been obtained (Doyle et al., 1998; Jiang et al., 2003a). What is it about the membrane spanning sequences of ion channels and the translocon that decides whether transmembrane regions traverse the membrane as they are synthesized or following assembly in the translocon? We do not know. Nonetheless, it is clear that the translocon participates actively in regulating nascent chain entry into the ER membrane (McCormick et al., 2003). Then there is the question of cytosolic loops. How do they get into the cytosol without collapsing the permeability barrier of the ER membrane? Two answers have been suggested. While the lumenal end of the translocon pore remains sealed, the ribosome-translocon junction could partially open to permit exit of cytosolic loops (Haigh and Johnson, 2002; Figure 1B, left). Alternatively, a 15 A˚ gap between the ribosome and the translocon, spanned by connectors (Figure 1B, right), could provide lateral openings through which cytosolic loops exit despite a colinear arrangement of the ribosome and translocon (Beckmann et al., 2001; Menetret et al., 2000). According to this model, a transmembrane segment would halt peptide translocation in the translocon and then integrate into the bilayer, dragging the next segment of nascent peptide through the lateral gap openings into the cytosol (Menetret et al., 2000). Folding along the Way Nascent channel peptides presumably fold during all stages of biogenesis. At each step, there are both constraints and aids. Some constraints are imposed by the ribosome (e.g., attachment to the PTC, distance from

the ribosome, associated factors [Fedorov and Baldwin, 1995; Makeyev et al., 1996]), some by the translocon. Likewise, these structures can serve as chaperones (Das et al., 1996; Kudlicki et al., 1997), along with soluble chaperones. There is ample room inside a translocon for a protein to fold (Kowarik et al., 2002). Indeed, folding can occur cotranslationally immediately after a domain has emerged from the ribosome while the rest of the peptide remains attached to the PTC only 40 to 50 amino acids away (Kowarik et al., 2002; Kudlicki et al., 1995; Makeyev et al., 1996; Nicola et al., 1999). There is also an intriguing possibility for tertiary structure formation inside the cramped ribosome tunnel. If the ribosome were conformationally dynamic, then it could pull and tug a malleable nascent peptide along through the tunnel from the PTC to the exit port while direct interactions between nascent peptide and ribosomal components promote folding, a kind of peristaltic remolding of the growing peptide on its journey. Alternatively, the tunnel could serve as a distensible conduit in which the peptide folds spontaneously (Hardesty and Kramer, 2001). Folding (when, where, how) is a complex issue for ion channels composed of many biogenic and functional domains. There are recognition domains, ligand binding domains, gates, sensors for gating, and pores. One naive premise is that these different domains fold independently and sequentially, N-terminal to C-terminal, as they are synthesized and exit the ribosome tunnel (Netzer and Hartl, 1998). Some domains may fold before targeting to the ER membrane, others at a later stage of biogenesis, perhaps requiring oligomerization. Some may enlist the aid of chaperones in the membrane or cytosol. We have been somewhat limited in assigning folding of specific domains to discrete stages in biogenesis, mostly due to a lack of tools to directly measure the folded state of the protein. Often, we infer a misfolded or unfolded state from the diagnosis that a channel protein is “trapped” in the ER, determined from measurements of surface labeling, immunohistochemistry, or lumi-

Review 269

nometry. These measurements tell us only about the location and nothing about the foldedness of the protein or a particular domain. Crosslinking assays have been useful in this respect, for example, to infer that steps of Kv folding and assembly alternate during channel biogenesis (Schulteis et al., 1998). Folding events to form the pore and voltage sensors may occur late in the sequence, after synthesis, insertion, tetramerization, and association with auxiliary subunits, but prior to formation of N- and C-terminal proximity. However, the spatial and temporal details of these folding events are not known. Direct measurements of folding of a monomeric T1 recognition domain (Figure 2A; for definition of “T1,” see below) in Kv channels has been made using an assay that relies on intramolecular crosslinking of pairs of cysteines engineered at the folded monomer T1 interface (Kosolapov and Deutsch, 2003). This recognition domain is largely folded soon after the Kv protein arrives at the ER membrane and only partially folded on the ribosome before targeting. These findings, along with those pertaining to tetramerization of this domain (see below) suggest that folding and tetramerization are coupled events that are facilitated by the ER membrane (Kosolapov and Deutsch, 2003; Lu et al., 2001). Similar experiments are necessary to elucidate the stepwise mechanism of folding of the voltage sensor, the permeation pore, and the gates. Last, but definitely not least, is the full association, folding, and integration of the complete quaternary channel protein in the bilayer. Only when mature proteinprotein, protein-lipid, and protein-aqueous interfaces are formed can the channel begin its career conducting ions across the membrane. The temporal sequence of interface formation is a yet-unexplored frontier. Finding the Right Partners Neurons contain a diverse repertoire of ion channel complexes, due largely to the combinatorics of pore-forming and auxiliary subunits. Thus, the cell must ensure that associating subunits coexist spatially and temporally, recognize each other, and provide sufficient stabilization energy to form a stable structure. When and where does this occur? With the exception of some gap junction channels (Sarma et al., 2002), the ER is the likely site of oligomerization for all ion channels (Deutsch, 2002). How does a cell maintain its channel repertoire and select among all possible combinations of poreforming subunit types and stoichiometries? The answer, in part, is encoded in the subunit protein itself, in socalled recognition domains. (However, we should remember that recognition domains may be different from stabilization domains that hold a mature channel together. In most cases, we have not yet experimentally discriminated these two.) We consider two examples: an N-terminal recognition domain, T1, in Kv channels, and a C-terminal leucine zipper, CLZ, in the A subunit of cyclic nucleotide-gated (CNG) channels. Only members of the same Kv1-Kv4 subfamilies coassemble to form channels (Covarrubias et al., 1991; Xu et al., 1995). This is due to a highly conserved sequence in the cytosolic N terminus of Kv channels, christened the “T1 domain” (first tetramerization domain [Li et al.,

1992; Shen et al., 1993]). Although T1-deleted Kv subunits associate promiscuously via transmembrane association domains to form stable, functional channels, both the rates and efficiency of tetramer formation are significantly lower (Deutsch, 2002). T1 domains of nascent Kv channel peptides interact and can be crosslinked prior to synthesis of the fulllength channel subunit (Lu et al., 2001). This intersubunit interaction is facilitated by the ER membrane, presumably due to concentration or restricted orientation of T1 domains and thus speedier oligomerization. Such membrane-dependent mechanisms could protect against premature folding and oligomerization of membrane proteins prior to targeting. T1-T1 interactions in nascent Kv subunits may serve to globally restrict subunits in the ER membrane, producing a local high concentration of subunits that kinetically expedites folding and tetramerization of transmembrane domains (Lu et al., 2001; Zerangue et al., 2000). A related issue is: what governs hetero- versus homotetramer formation for the same recognition domain? Heteromeric Kv channels with unique properties exist in most tissues, including brain and heart, and are critical for normal function. The principles determining whether channel subunits assemble predominantly with themselves or with other members of the gene family are thus of great importance but as yet undefined. When two different monomers containing compatible recognition domains are expressed within the same cell, the resultant compositions and stoichiometries of tetramers could vary extensively. At one extreme, preferential association yields only homotetramers; at the other, random association yields a binominally weighted mixture of homo- and heterotetramers. However, in some cases, only heterotetramers form (Deutsch, 2002). Although it appears that the degree of temporal overlap and kinetics of subunit expression govern the proportions of homo- and heterotetramers (Panyi and Deutsch, 1996), little is known (with the exception of CNG channels [Zhong et al., 2003], below) about the mechanisms for this regulated distribution in vivo. One fascinating possibility is that different but related mRNAs are bound in a translational unit via RNA binding proteins and are coordinately translated (Keene and Tenenbaum, 2002). Not all recognition domains are located exclusively in the N terminus of K⫹ channels. Rather, some appear to be in the C terminus or transmembrane domains (see Deutsch, 2002). For example, CNG channels are tetrameric, with a stoichiometry of 3A:1B subunits (Weitz et al., 2002; Zheng et al., 2002; Zhong et al., 2002). A CLZ domain, found only in the C terminus of the A subunit, is responsible for this selective heterotetramer formation due to CLZ trimerization of the A subunits (Zhong et al., 2003). However, if C-terminal recognition domains are responsible for tetramer formation and are also the last to be synthesized, then what is the mechanism for oligomerization? The mechanism is likely to be distinctly different from that involving N-terminal recognition domains, which are translated and functionally available prior to synthesis of the rest of the nascent subunit. Possibly the channel is held in the translocon and released into the bilayer in a coupled, synchronous event with C-terminal tetramerization. Nothing is known about the temporal or spatial sequence of events of this tetra-

Neuron 270

merization with respect to cotranslation and/or integration. In principle, any peptide domain imbued with the necessary properties (e.g., high affinity, specificity) could serve as a recognition and/or stabilization domain. Proof of principle was elegantly demonstrated by Zerangue et al. (2000), who replaced the T1 domain of Shaker B with a coiled-coil sequence (GCN4-L1) that forms parallel tetramers. This artificial tetramerization domain restored efficient channel formation to a T1-deleted Shaker protein. How do four Kv subunits come together? Tetramers could form via concerted or stepwise mechanisms. Two stepwise possibilities were explored by Tu and Deutsch (1999) for Kv1.3. One is by sequential addition of monomers to form dimers, then trimers, and finally, tetramers. Alternatively, monomers may associate to form dimers, which then dimerize to form tetramers. The dominant pathway in Kv tetramer formation is dimerization of dimers. Moreover, conformational changes accompany dimer formation, thereby creating new interaction sites that are used in the next stage to form tetramers. The specific interactions underlying each stage are unknown. The Social Scene A mature ion channel in the plasma membrane is more than just pore-forming subunits. It is a multicomponent macrocomplex of pore-forming subunits and other auxiliary proteins that modulate function. Thus, a critical element of ion channel ontogeny is the association of these components to ultimately form the mature channel complex. In principle, all compartments along the ER-toplasma membrane journey could host such association. Auxiliary proteins have been crafted to serve dual roles: modulators of function and modulators of biogenesis/ trafficking. We refer to this as a cell’s economy of purpose. Except in a few cases (see below), the mechanisms of biogenic facilitation are not known, nor where, when, or how auxiliary subunits associate with their pore-forming counterparts. Association could be cotranslational or posttranslational. Auxiliary membrane proteins could associate with pore-forming subunits at their mutual peripheries or could intercalate between pore-forming subunits. One stoichiometry or multiple stoichiometries may predominate. While these fundamental assembly issues remain both intriguing and in most cases, unresolved, they hold clues to the role of auxiliary subunits in assembly. We will consider only a couple of examples. First, there is calmodulin (CaM), a ubiquitous cytosolic Ca binding protein. It associates with many channel types, including the Ca2⫹-activated small conductance K⫹ channel (SK) (Figure 4). CaM, which constitutively associates with SK subunits in a calcium-independent manner (Xia et al., 1998), binds calcium, which activates the channel. When does this marriage occur? Apparently early in biogenesis, because surface expression of SK requires CaM (but not Ca2⫹ bound CaM [Joiner et al., 2001; Lee et al., 2003]). The Ca-sensing function of CaM does not seem to be involved in its biogenic role. Another example is the L-type Ca channel, which also constitutively binds CaM. Surface expression is governed by Ca-

dependent CaM (Peterson et al., 1999). These examples illustrate an escort role and perhaps even a chaperone or scaffolding role, for the auxiliary protein. However, CaM modulation of ion channel biogenesis and function is idiosyncratic. For KCNQ potassium channels, CaM is also constitutively associated but inhibition of CaM binding does not prevent surface expression (Wen and Levitan, 2002). Dual roles may be ascribed to another class of cytosolic auxiliary subunits, the so-called ␤ subunits that associate with Kv channels (“␣ subunits”). These include soluble cytosolic Kv␤ isoforms and K-channel interacting protein (KChIP) subunits. (KChIP is also a calcium binding protein, but neither its functional role nor biogenic role has been shown to be Ca dependent.) For some of these, we know something about the specificity and association domains in the pore-forming subunit and in the auxiliary subunit (Bahring et al., 2001; Xu et al., 1998). While these auxiliary subunits have marked functional effects, particularly on gating, they also increase stability and cell surface expression of Kv channels (An et al., 2000; Bahring et al., 2001; Beck et al., 2002; Xu et al., 1998). And another, completely different auxiliary subunit of the Kv4 family, related to dipeptidases that span the membrane and contribute to cell adhesion, modulates gating and surface expression (Nadal et al., 2003). ␣ and ␤ subunits associate in the ER (Nagaya and Papazian, 1997), and a chaperone-like role early in assembly in the ER has been proposed for ␤ subunits (Accili et al., 1997; Shi et al., 1996). The structure of the complexes of the pore-forming channel and ␤ subunits consists of the T1 tetramer (see above) docked onto the ␤ tetramer to form an octomeric structure with a stoichiometry of T14␤4 (Figure 4; Gulbis et al., 2000). However, we know nothing about the sequence of events that gives rise to this structure. Although the ␣:␤ stoichiometry first reported was 1:1, it now appears that Kv␤1 and Kv␤2 associate with ␣ subunits in two different functional stoichiometries: ␣4␤1n (where n ⫽ 0–4) and ␣4␤24, which may underlie their different roles in channel function and expression, respectively (Xu et al., 1998). How do these different stoichiometries arise? Do ␤ subunits first self-associate to form various oligomeric ␤n stoichiometries and then associate with an ␣ tetramer; do ␤ subunits individually, and sequentially, associate with ␣ tetramers; or do ␣ and ␤ subunits associate to form a heterodimer, which subsequently associates with other ␣␤ dimers to form the octomeric 4:4 complex? The first scenario is likely to prevail in the case of ␣4␤24, consistent with the positive cooperativity for ␤2 binding to itself, and would account for the apparent higher avidity of ␤2 for Kv channels, since a tetrameric ␤2 would provide multivalent interactions (Xu et al., 1998). It is the sequence of association events that may shed light on the role of auxiliary subunits in assembly (Zerangue et al., 2000). For example, if T1 self-association occurs first, T1 could serve as a scaffold for folding the rest of the channel structure. Likewise, Kv␤ tetramer, if it associates early in assembly, could scaffold the tetramerization of T1. Even the reverse may be considered: the early tetramerization of the folded T1 domain may facilitate Kv␤ tetramerization. Kv␤’s possible role as chaperone, scaffold, modulator, and/or transducer

Review 271

Figure 4. Macrocomplexes of Ion Channels Kv, SK, and voltage-dependent calcium channels (Cav) assemble with auxiliary subunits (dark blue). For Kv channels, the T1 domains associate with ␤ subunits (foreshortened cylinder; Gulbis et al., 2000); for SK channels, the C terminus associates with calmodulin (two-lobed symbol; Lee et al., 2003); for Cav, both N and C termini and the loop between domains II and III associate with AKAP, which recruits kinase (PKA) and phosphatase (PP2A) (Altier et al., 2002; Davare et al., 1999). In all cases, the structures are not drawn to scale. For simplicity, the N termini for SK are not shown, nor are the C termini for the Kv channel.

will likely be implicated by the temporal sequence and stoichiometry of ␣␤ association. Finally, we must again ask how the right partners find each other. In this case, the partners are the constitutively associated auxiliary subunits and pore-forming subunits. If association is cotranslational, then one provocative possibility is the one described above for heteromultimer formation. mRNAs for similarly fated subunits that form a common structure could be prepackaged together in ribonucleoprotein complexes, analogous to operons and their polycistronic mRNAs in prokaryotic organisms, and coordinately translated. Such clustered complexes do exist for some mammalian transcripts (Keene and Tenenbaum, 2002). Stop-and-Go Traffic Neuronal excitability requires more than the simple manufacture and assembly of ion channel proteins. The cell must provide mechanisms for appropriate exit from the ER, targeting, localization, abundance, and turnover, all of which govern the levels of surface expression and distribution of ion conductances. Several mechanisms underlie the ER exit decision for an ion channel. First, there are retention/retrieval signals for improperly folded or assembled channels, e.g., CFTR, KATP, GABAB, AChR, and NMDA channels (Margeta-Mitrovic et al., 2000; Scott et al., 2001; Standley et al., 2000; Wang et al., 2002a; Xia et al., 2001; Zerangue et al., 1999). Second, anterograde signals may be necessary for exporting properly assembled channels, e.g., KIR, Kv1.4, and SUR1 (Li et al., 2000; Ma et al., 2001, 2002; Sharma et al., 1999; Stockklausner et al., 2001). Third, the phosphorylation state of the channel could determine exit from the ER (Scott et al., 2001; Xia et al., 2001). Fourth, association of the channel with scaffolding/anchoring proteins (Scott et al., 2001; Standley et al., 2000; Tiffany et al., 2000; Xia et al., 2001) may be required for proper surface expression. Although the mechanism is unknown, scaffold binding could influence the ability of retention and anterograde signals to do their jobs. Fifth, subunit composition and stoichiometry determine surface expression. Sixth, misfolding of the protein can prohibit ER exit, even in the absence of known retention signals. Here, we will focus on only a few of these cases, again illustrating the principle of economy of purpose, i.e., the cell often uses the same auxiliary subunit for both assembly and function.

A good example of primary sequence motifs directing retrograde movement is the KATP channel, which is assembled as an octomer with a 1:1 stoichiometry of potassium inward rectifying channel subunits (KIR) and auxiliary SUR subunits (sulfonyl urea receptor). The SUR subunits modulate function and drug sensitivity of channel activity (Babenko et al., 1998). Assembly of KIR and SUR occurs in the ER and early Golgi (Zerangue et al., 1999), and putative recognition domains have been identified (Schwappach et al., 2000; Tucker and Ashcroft, 1999). Both the KIR and SUR subunits have cytoplasmic RKR sequences, which must be masked by assembly of the octomer before the channel can be transported to the cell surface (Zerangue et al., 1999). Similar motifs exist in a cytosolic domain of CFTR (RXT; Gilbert et al., 1998) and in the C terminus of a GABAB receptor subunit (RSRR; Margeta-Mitrovic et al., 2000). In the latter case, heterodimeric coiled-coil interactions shield the motif, allowing the complex to be expressed at the cell surface. Masking of retention/retrieval motifs could be a general mechanism for quality control in ion channel assembly. A very different motif, PL(Y/F)(F/Y)xxN, in the first transmembrane domain (M1) of the nicotinic acetylcholine receptor channel (nAChR) ␣ subunit, governs trafficking of correctly assembled pentameric channels. This motif must be masked for forward trafficking to occur, thus preventing surface expression of unassembled monomers or partially assembled complexes. Moreover, this motif, if exposed, promotes rapid degradation of the protein (Wang et al., 2002a). Yet another example of masking a retention motif is in the CNG channels. A distal C-terminal domain in the A subunit interacts with an N-terminal region of the B subunit to mask a short sequence in the B subunit, thereby allowing surface expression of the CNG channel (Trudeau and Zagotta, 2002). Disruption of this regulatory interaction by a truncation mutation underlies retinitis pigmentosa, an inherited eye disease. The second class of exit determinants, anterograde signals, is illustrated by KIR 1.1 and 2.1 channels. Here, diacidic motifs (e.g., ENE, EXD) function as direct ER export signals (Ma et al., 2001). These signals do not act by promoting correct folding or assembly, nor are they context dependent. When inserted into Kv1.2, a poor surface expressor (Shi et al., 1996), FCYENE promoted Kv1.2 expression, similar to the effects produced by Kv␤2. This raises the intriguing possibility that Kv␤2

Neuron 272

may facilitate surface expression because it forms a tight complex with ␣ subunits and contains export signals. This would suggest a speedy escort function rather than a classical chaperone folding function for Kv␤2. The flanking residues of diacidic motifs are also critical for ER export and vary among K⫹ channels, thus providing a mechanism for differential distributions of K⫹ channels on the cell surface. A combination of both retrograde and anterograde signals can together govern ER exit. For example, C-terminal export signals also exist in the SUR subunit and regulate KATP surface expression (Sharma et al., 1999). Together with the retrograde RKR motif, these anterograde signals ensure that correctly assembled octomeric KATP arrives at the plasma membrane. Another example is KCNK, a homodimeric 2-pore K⫹ channel that modulates excitability of nerves and muscle (Goldstein et al., 2001; Lesage and Lazdunski, 2000). In this case, a third determinant, phosphorylation, joins in the regulation of ER exit. Dibasic motifs (e.g., one, two, or three sequential arginines or lysines) lead to ER retention (Nilsson et al., 1989; Teasdale and Jackson, 1996; Zerangue et al., 2001), and a cytosolic protein, 14-3-3, allows forward movement. The unphosphorylated channel supports the retention mechanism; the phosphorylated channel supports the forward mechanism and inhibits the retention mechanism (O’Kelly et al., 2002). This scenario is likely to be quite general. For example, class II-invariant chain complexes have a dibasic retention motif, and anterograde trafficking involves a similar push-pull, phosphorylation-dependent mechanism (Anderson et al., 1999; Kuwana et al., 1998; O’Kelly et al., 2002; Schutze et al., 1994). In these cases, a diversity of serine kinase phosphorylation sites regulates trafficking. No doubt, sites for many other kinases will play a role as well. Similarly, certain nicotinic AChR channels are retained in the ER due to a dibasic motif in a cytosolic loop of the channel protein. 14-3-3 releases the channel in a phosphorylation-dependent interaction to forward transport (Jeanclos et al., 2001; Keller et al., 2001; O’Kelly et al., 2002). Phosphorylation is also implicated in the biogenic role of Kv␤ subunits. These subunits have many phosphorylation sites and cytosolic ZIP1 and ZIP2 (PKC-␨ interacting protein) bind to Kv␤2 subunits and PKC-␨, thus serving as physical linkers (Gong et al., 1999). ZIP1 and ZIP2 differentially stimulate phosphorylation of Kv␤2 by PKC-␨ and may be responsible for specific regulation of PKC-␨ phosphorylation. When does ZIP stimulate PKC-␨ phosphorylation of Kv␤2? Could it be before Kv␣ and Kv␤2 associate, or only after the octomeric complex is assembled? If before, then phosphorylation may play a biogenic role. A fourth determinant of ER exit is the association of channels with scaffolding/anchoring proteins. Such proteins recruit others to the channel to form macrocomplexes. Included among the ion channel macrocomplexes at the plasma membrane are direct channelkinase-phosphatase complexes, which provide efficient regulation of channel function. Some of these complexes are mediated by anchoring proteins, e.g., A-kinase anchoring proteins (AKAPs), which recruit protein kinase A (PKA) to the channel (Gray et al., 1998). Phosphorylation then modulates channel activity. However, AKAP has a dual mission: it also regulates surface

expression. In L-type calcium channels, AKAP regulates expression independently of PKA activation. The mechanism appears similar to the masking strategy used in KATP channel assembly. AKAP likely masks a retention/ retrieval motif in the II-III linker of the Ca2⫹ channel (Altier et al., 2002) (Figure 4). This raises the intriguing possibility that scaffolding proteins at the plasma membrane, used to complex the L-type Ca2⫹ channels with PKA and PP2A (Davare et al., 1999), may also facilitate channel biogenesis. A similar role has been ascribed to the calcium channel ␤ subunit, which masks a retention motif in the I-II linker domain of the ␣ subunit (Bichet et al., 2000). In addition to a masking mechanism, ␤ binding could induce conformational modifications of the channel structure, thus permitting it to pass ER inspection and exit. Alternatively, the ␤ subunit itself, by binding to the I-II loop, could be conformationally altered and thus acquire anterograde signaling potential. If large complexes of pore-forming and regulatory auxiliary units exist in the ER, then this would provide a panoply of interactions for cellular crosstalk. Specifically, signaling pathways might regulate biogenic events. For example, ZIP proteins, which themselves can be regulated by growth factors, link PKC-␨ with Kv␤ subunits (Gong et al., 1999), which, in turn, have chaperone, scaffolding, and/or export roles in Kv assembly. Another example is G␤␥, which binds to Kv1.1 and to Kv␤1.1 and stabilizes the Kv1.1-Kv␤1.1 complex, possibly early in biosynthesis in the ER (Jing et al., 1999). This provides the tantalizing possibility of coupling G protein signal transduction, initiated by extracellular signals at the cell surface, to control of channel assembly in biosynthetic intracellular organelles. By extension, we imagine that the combinations of other signaling molecules, auxiliary subunits and pore-forming channel subunits might govern channel biogenesis. Finally, if macrocomplexes that recruit kinases and phosphatases are assembled in the ER, then phosphorylation/dephosphorylation could also modulate biogenesis through signaling pathways. Another determinant of ER exit, subunit composition and stoichiometry, may influence exit as a consequence of masking defaults. However, in most cases, putative retrograde and anterograde motifs are unknown. This is true for so-called silent subunits that become functional if expressed with specific Kv channel subunits (Ficker et al., 2001; Kennedy et al., 1999; Ottschytsch et al., 2002; Zarei et al., 2001). Alternatively, association of other combinations can prevent expression at the cell surface (Bond et al., 1994; Pessia et al., 1996; Tucker et al., 1996). Both mechanisms conspire to modulate surface conductance (Manganas and Trimmer, 2000). This case is particularly illustrated by the Kv channel families. The Kv channel family includes eleven different subfamilies, Kv1.-Kv11. The first four subfamilies form functional channels when expressed alone, whereas Kv5.Kv11. do not (Ottschytsch et al., 2002). Some of the latter Kv channels contain no known retention or export sequences, suggesting that the defect is an assembly problem. These silent subunits will not self-associate. If oligomerization and folding are coupled events (Kosolapov and Deutsch, 2003; see above), then inhibited selfassociation also precludes complete folding and leads

Review 273

to ER retention. However, these silent subunits will form functional heterotetrameric channels with members of the Kv2 family, producing currents with different properties from homotetramer Kv2 currents (Ottschytsch et al., 2002; Patel et al., 1997; Salinas et al., 1997; Zhu et al., 1999). The silent subunits have wide tissue distribution, including the brain, and are retained in the ER (Ottschytsch et al., 2002). Because they represent approximately onethird of all Kv subunits, silent subunits provide a large pool of modifiers from which the cell may choose, thus significantly molding the repertoire of K⫹ conductances at the cell surface. And, finally, a sixth determinant of ER exit is the folded state of the channel. For example, several mutations in the cardiac HERG channel and in the CFTR channel fail to exit the ER (Ficker et al., 2000; Furutani et al., 1999; Skach, 2000; Zhou et al., 1998) and the defect is usually assumed to be misfolding. However, what does this mean? Does it mean minimal disruption that exposes a retention motif, a global unwinding of a helix, or an interrupted intramolecular interface? Moreover, a misfolding diagnosis could be erroneous, because aberrant trafficking may have other causes. In some cases, defective trafficking can be restored by lowering incubation temperatures or by channel blockers or chemical chaperones (Howard and Welch, 2002; Paulussen et al., 2002; Rajamani et al., 2002; Zhou et al., 1999). These examples may offer clues. Restoration with channel blockers, for example, suggests a folding defect and that the blocker acts as a stabilizing ligand.

kinase C regulates MHC class II trafficking to antigen-processing compartments. J. Immunol. 163, 5435–5443.

A Final Comment We have seen that the birth of a channel proceeds in a complex and wondrous fashion, from protein formation within the ribosome to macrocomplex assembly in the ER. Though this ontological journey has many unknown steps, the questions raised here are but mere hints of the awesome answers to come.

Choi, K.M., and Brimacombe, R. (1998). The path of the growing peptide chain through the 23S rRNA in the 50S ribosomal subunit; a comparative cross-linking study with three different peptide families. Nucleic Acids Res. 26, 887–895.

Acknowledgments Thanks to R. Horn, A.E. Johnson, C. Nicchitta, and M. Covarrubias for critical reading of the manuscript; and special thanks to A.E. Johnson for helpful discussions. References Accili, E.A., Kiehn, J., Yang, Q., Wang, Z., Brown, A.M., and Wible, B.A. (1997). Separable Kvbeta subunit domains alter expression and gating of potassium channels. J. Biol. Chem. 272, 25824–25831. Altier, C., Dubel, S.J., Barrere, C., Jarvis, S.E., Stotz, S.C., Spaetgens, R.L., Scott, J.D., Cornet, V., De Waard, M., Zamponi, G.W., et al. (2002). Trafficking of L-type calcium channels mediated by the postsynaptic scaffolding protein AKAP79. J. Biol. Chem. 277, 33598–33603. An, W.F., Bowlby, M.R., Betty, M., Cao, J., Ling, H.P., Mendoza, G., Hinson, J.W., Mattsson, K.I., Strassle, B.W., Trimmer, J.S., and Rhodes, K.J. (2000). Modulation of A-type potassium channels by a family of calcium sensors. Nature 403, 553–556. Anderson, D.J., Walter, P., and Blobel, G. (1982). Signal recognition protein is required for the integration of acetylcholine receptor delta subunit, a transmembrane glycoprotein, into the endoplasmic reticulum membrane. J. Cell Biol. 93, 501–506. Anderson, H.A., Bergstralh, D.T., Kawamura, T., Blauvelt, A., and Roche, P.A. (1999). Phosphorylation of the invariant chain by protein

Babenko, A.P., Aguilar-Bryan, L., and Bryan, J. (1998). A view of SUR/KIR6.X, KATP channels. Annu. Rev. Physiol. 60, 667–687. Bahring, R., Dannenberg, J., Peters, H.C., Leicher, T., Pongs, O., and Isbrandt, D. (2001). Conserved Kv4 N-terminal domain critical for effects of Kv channel-interacting protein 2.2 on channel expression and gating. J. Biol. Chem. 276, 23888–23894. Ban, N., Nissen, P., Hansen, J., Moore, P.B., and Steitz, T.A. (2000). The complete atomic structure of the large ribosomal subunit at 2.4 A resolution. Science 289, 905–920. Beck, E.J., Bowlby, M., An, W.F., Rhodes, K.J., and Covarrubias, M. (2002). Remodelling inactivation gating of Kv4 channels by KChIP1, a small-molecular-weight calcium-binding protein. J. Physiol. 538, 691–706. Beckmann, R., Spahn, C.M., Eswar, N., Helmers, J., Penczek, P.A., Sali, A., Frank, J., and Blobel, G. (2001). Architecture of the proteinconducting channel associated with the translating 80S ribosome. Cell 107, 361–372. Berisio, R., Schluenzen, F., Harms, J., Bashan, A., Auerbach, T., Baram, D., and Yonath, A. (2003). Structural insight into the role of the ribosomal tunnel in cellular regulation. Nat. Struct. Biol. 10, 366–370. Bichet, D., Cornet, V., Geib, S., Carlier, E., Volsen, S., Hoshi, T., Mori, Y., and De Waard, M. (2000). The I–II loop of the Ca2⫹ channel alpha1 subunit contains an endoplasmic reticulum retention signal antagonized by the beta subunit. Neuron 25, 177–190. Bond, C.T., Pessia, M., Xia, X.M., Lagrutta, A., Kavanaugh, M.P., and Adelman, J.P. (1994). Cloning and expression of a family of inward rectifier potassium channels. Receptors Channels 2, 183–191. Borel, A.C., and Simon, S.M. (1996). Biogenesis of polytopic membrane proteins: membrane segments assemble within translocation channels prior to membrane integration. Cell 85, 379–389.

Clark, P.L., and King, J. (2001). A newly synthesized, ribosomebound polypeptide chain adopts conformations dissimilar from early in vitro refolding intermediates. J. Biol. Chem. 276, 25411–25420. Covarrubias, M., Wei, A.A., and Salkoff, L. (1991). Shaker, Shal, Shab, and Shaw express independent K⫹ current systems. Neuron 7, 763–773. Crowley, K.S., Liao, S., Worrell, V.E., Reinhart, G.D., and Johnson, A.E. (1994). Secretory proteins move through the endoplasmic reticulum membrane via an aqueous, gated pore. Cell 78, 461–471. Das, B., Chattopadhyay, S., Bera, A.K., and Dasgupta, C. (1996). In vitro protein folding by ribosomes from Escherichia coli, wheat germ and rat liver: the role of the 50S particle and its 23S rRNA. Eur. J. Biochem. 235, 613–621. Davare, M.A., Dong, F., Rubin, C.S., and Hell, J.W. (1999). The A-kinase anchor protein MAP2B and cAMP-dependent protein kinase are associated with class C L-type calcium channels in neurons. J. Biol. Chem. 274, 30280–30287. Deutsch, C. (2002). Potassium channel ontogeny. Annu. Rev. Physiol. 64, 19–46. Do, H., Falcone, D., Lin, J., Andrews, D.W., and Johnson, A.E. (1996). The cotranslational integration of membrane proteins into the phospholipid bilayer is a multistep process. Cell 85, 369–378. Doyle, D.A., Cabral, J.M., Pfuetzner, R.A., Kuo, A., Gulbis, J.M., Cohen, S.L., Chait, B.T., and MacKinnon, R. (1998). The structure of the potassium channel: Molecular basis of K⫹ conduction and selectivity. Science 280, 69–76. Fedorov, A.N., and Baldwin, T.O. (1995). Contribution of cotranslational folding to the rate of formation of native protein structure. Proc. Natl. Acad. Sci. USA 92, 1227–1231. Ficker, E., Thomas, D., Viswanathan, P.C., Dennis, A.T., Priori, S.G., Napolitano, C., Memmi, M., Wible, B.A., Kaufman, E.S., Iyengar, S.,

Neuron 274

et al. (2000). Novel characteristics of a misprocessed mutant HERG channel linked to hereditary long QT syndrome. Am. J. Physiol. Heart Circ. Physiol. 279, H1748–H1756.

Joiner, W.J., Khanna, R., Schlichter, L.C., and Kaczmarek, L.K. (2001). Calmodulin regulates assembly and trafficking of SK4/IK1 Ca2⫹-activated K⫹ channels. J. Biol. Chem. 276, 37980–37985.

Ficker, E.K., Dennis, A.T., Obejero-Paz, C.A., Castaldo, P., Taglialatela, M., and Brown, A.M. (2001). Retention in the endoplasmic reticulum as a mechanism of dominant-negative current suppression in human long QT syndrome. Biophys. J. 80, 215a.

Keene, J.D., and Tenenbaum, S.A. (2002). Eukaryotic mRNPs may represent posttranscriptional operons. Mol. Cell 9, 1161–1167.

Furutani, M., Trudeau, M.C., Hagiwara, N., Seki, A., Gong, Q., Zhou, Z., Imamura, S., Nagashima, H., Kasanuki, H., Takao, A., et al. (1999). Novel mechanism associated with an inherited cardiac arrhythmia: defective protein trafficking by the mutant HERG (G601S) potassium channel. Circulation 99, 2290–2294. Gabashvili, I.S., Gregory, S.T., Valle, M., Grassucci, R., Worbs, M., Wahl, M.C., Dahlberg, A.E., and Frank, J. (2001). The polypeptide tunnel system in the ribosome and its gating in erythromycin resistance mutants of L4 and L22. Mol. Cell 8, 181–188. Gilbert, A., Jadot, M., Leontieva, E., Wattiaux-De, C.S., and Wattiaux, R. (1998). Delta F508 CFTR localizes in the endoplasmic reticulumGolgi intermediate compartment in cystic fibrosis cells. Exp. Cell Res. 242, 144–152. Goldstein, S.A., Bockenhauer, D., O’Kelly, I., and Zilberberg, N. (2001). Potassium leak channels and the KCNK family of twoP-domain subunits. Nat. Rev. Neurosci. 2, 175–184.

Keller, S.H., Lindstrom, J., Ellisman, M., and Taylor, P. (2001). Adjacent basic amino acid residues recognized by the COP I complex and ubiquitination govern endoplasmic reticulum to cell surface trafficking of the nicotinic acetylcholine receptor alpha-Subunit. J. Biol. Chem. 276, 18384–18391. Kennedy, M.E., Nemec, J., Corey, S., Wickman, K., and Clapham, D.E. (1999). GIRK4 confers appropriate processing and cell surface localization to G-protein-gated potassium channels. J. Biol. Chem. 274, 2571–2582. Komar, A.A., Kommer, A., Krasheninnikov, I.A., and Spirin, A.S. (1997). Cotranslational folding of globin. J. Biol. Chem. 272, 10646– 10651. Kosolapov, A., and Deutsch, C. (2003). Folding of the voltage-gated K⫹ channel T1 recognition domain. J. Biol. Chem. 278, 4305–4313. Kowarik, M., Kung, S., Martoglio, B., and Helenius, A. (2002). Protein folding during cotranslational translocation in the endoplasmic reticulum. Mol. Cell 10, 769–778.

Gong, J., Xu, J., Bezanilla, M., van Huizen, R., Derin, R., and Li, M. (1999). Differential stimulation of PKC phosphorylation of potassium channels by ZIP1 and ZIP2. Science 285, 1565–1569.

Kramer, G., Kudlicki, W., McCarthy, D., Tsalkova, T., Simmons, D., and Hardesty, B. (1999). N-terminal and C-terminal modifications affect folding, release from the ribosomes and stability of in vitro synthesized proteins. Int. J. Biochem. Cell Biol. 31, 231–241.

Gray, P.C., Scott, J.D., and Catterall, W.A. (1998). Regulation of ion channels by cAMP-dependent protein kinase and A-kinase anchoring proteins. Curr. Opin. Neurobiol. 8, 330–334.

Kudlicki, W., Chirgwin, J., Kramer, G., and Hardesty, B. (1995). Folding of an enzyme into an active conformation while bound as peptidyl-tRNA to the ribosome. Biochemistry 34, 14284–14287.

Gulbis, J.M., Zhou, M., Mann, S., and MacKinnon, R. (2000). Structure of the cytoplasmic beta subunit-T1 assembly of voltage-dependent K⫹ channels. Science 289, 123–127.

Kudlicki, W., Coffman, A., Kramer, G., and Hardesty, B. (1997). Ribosomes and ribosomal RNA as chaperones for folding of proteins. Fold. Des. 2, 101–108.

Haigh, N.G., and Johnson, A.E. (2002). A new role for BiP: closing the aqueous translocon pore during protein integration into the ER membrane. J. Cell Biol. 156, 261–270.

Kuwana, T., Peterson, P.A., and Karlsson, L. (1998). Exit of major histocompatibility complex class II-invariant chain p35 complexes from the endoplasmic reticulum is modulated by phosphorylation. Proc. Natl. Acad. Sci. USA 95, 1056–1061.

Hamman, B.D., Chen, J.C., Johnson, E.E., and Johnson, A.E. (1997). The aqueous pore through the translocon has a diameter of 40–60 A during cotranslational protein translocation at the ER membrane. Cell 89, 535–544. Hardesty, B., and Kramer, G. (2001). Folding of a nascent peptide on the ribosome. Prog. Nucleic Acid Res. Mol. Biol. 66, 41–66. Harms, J., Schluenzen, F., Zarivach, R., Bashan, A., Gat, S., Agmon, I., Bartels, H., Franceschi, F., and Yonath, A. (2001). High resolution structure of the large ribosomal subunit from a mesophilic eubacterium. Cell 107, 679–688. Hegde, R.S., and Lingappa, V.R. (1997). Membrane protein biogenesis: regulated complexity at the endoplasmic reticulum. Cell 91, 575–582. Howard, M., and Welch, W.J. (2002). Manipulating the folding pathway of delta F508 CFTR using chemical chaperones. Methods Mol. Med. 70, 267–275. Jeanclos, E.M., Lin, L., Treuil, M.W., Rao, J., DeCoster, M.A., and Anand, R. (2001). The chaperone protein 14–3-3eta interacts with the nicotinic acetylcholine receptor alpha 4 subunit. Evidence for a dynamic role in subunit stabilization. J. Biol. Chem. 276, 28281– 28290. Jiang, Y., Lee, A., Chen, J., Ruta, V., Cadene, M., Chait, B.T., and MacKinnon, R. (2003a). X-ray structure of a voltage-dependent K⫹ channel. Nature 423, 33–41. Jiang, Y., Ruta, V., Chen, J., Lee, A., and MacKinnon, R. (2003b). The principle of gating charge movement in a voltage-dependent K⫹ channel. Nature 423, 42–48. Jing, J., Chikvashvili, D., Singer-Lahat, D., Thornhill, W.B., Reuveny, E., and Lotan, I. (1999). Fast inactivation of a brain K⫹ channel composed of Kv1.1 and Kvbeta1.1 subunits modulated by G protein beta gamma subunits. EMBO J. 18, 1245–1256. Johnson, A.E., and van Waes, M.A. (1999). The translocon: a dynamic gateway at the ER membrane. Annu. Rev. Cell Dev. Biol. 15, 799–842.

Lee, W.-S., Ngo-Anh, T.J., Bruening-Wright, A., Maylie, J., and Adelman, J.P. (2003). Small conductance Ca2⫹-activated K⫹ channels and calmodulin: Cell surface expression and gating. J. Biol. Chem., in press. Lesage, F., and Lazdunski, M. (2000). Molecular and functional properties of two-pore-domain potassium channels. Am. J. Physiol. Renal Fluid Electrol. Physiol. 279, F793–F801. Li, M., Jan, Y.N., and Jan, L.Y. (1992). Specification of subunit assembly by the hydrophilic amino-terminal domain of the Shaker potassium channels. Science 257, 1225–1230. Li, D., Takimoto, K., and Levitan, E.S. (2000). Surface expression of Kv1 channels is governed by a C-terminal motif. J. Biol. Chem. 275, 11597–11602. Liao, S., Lin, J., Do, H., and Johnson, A.E. (1997). Both lumenal and cytosolic gating of the aqueous ER translocon pore are regulated from inside the ribosome during membrane protein integration. Cell 90, 31–41. Lovett, P.S., and Rogers, E.J. (1996). Ribosome regulation by the nascent peptide. Microbiol. Rev. 60, 366–385. Lu, Y., Xiong, X., Helm, A., Kimani, K., Bragin, A., and Skach, W.R. (1998). Co- and posttranslational translocation mechanisms direct cystic fibrosis transmembrane conductance regulator N terminus transmembrane assembly. J. Biol. Chem. 273, 568–576. Lu, Y., Turnbull, I.R., Bragin, A., Carveth, K., Verkman, A.S., and Skach, W.R. (2000). Reorientation of aquaporin-1 topology during maturation in the endoplasmic reticulum. Mol. Biol. Cell 11, 2973– 2985. Lu, J., Robinson, M.R., Edwards, D., and Deutsch, C. (2001). T1–T1 interactions occur in ER membranes while nascent Kv peptides are still attached to ribosomes. Biochemistry 40, 10934–10946. Ma, D., Zerangue, N., Lin, Y.F., Collins, A., Yu, M., Jan, Y.N., and Jan, L.Y. (2001). Role of ER export signals in controlling surface potassium channel numbers. Science 291, 316–319.

Review 275

Ma, D., Zerangue, N., Raab-Graham, K., Fried, S.R., Jan, Y.N., and Jan, L.Y. (2002). Diverse trafficking patterns due to multiple traffic motifs in G protein-activated inwardly rectifying potassium channels from brain and heart. Neuron 33, 715–729. Makeyev, E.V., Kolb, V.A., and Spirin, A.S. (1996). Enzymatic activity of the ribosome-bound nascent polypeptide. FEBS Lett. 378, 166–170. Manganas, L.N., and Trimmer, J.S. (2000). Subunit composition determines Kv1 potassium channel surface expression. J. Biol. Chem. 275, 29685–29693. Margeta-Mitrovic, M., Jan, Y.N., and Jan, L.Y. (2000). A trafficking checkpoint controls GABA(B) receptor heterodimerization. Neuron 27, 97–106. Martoglio, B., Hofmann, M.W., Brunner, J., and Dobberstein, B. (1995). The protein-conducting channel in the membrane of the endoplasmic reticulum is open laterally toward the lipid bilayer. Cell 81, 207–214. McCormick, P.J., Miao, Y., Shao, Y., Lin, J., and Johnson, A.E. (2003). Co-translational protein integration into the ER membrane is mediated by the binding of nascent chains to translocon proteins. Mol. Cell 12, 329–341. Meacock, S.L., Lecomte, F.J.L., Crawshaw, S.G., and High, S. (2002). Different transmembrane domains associate with distinct endoplasmic reticulum components during membrane integration of a polytopic protein. Mol. Biol. Cell 13, 4114–4129. Menetret, J.F., Neuhof, A., Morgan, D.G., Plath, K., Radermacher, M., Rapoport, T.A., and Akey, C.W. (2000). The structure of ribosomechannel complexes engaged in protein translocation. Mol. Cell 6, 1219–1232. Morris, D.R., and Geballe, A.P. (2000). Upstream open reading frames as regulators of mRNA translation. Mol. Cell. Biol. 20, 8635– 8642. Mothes, W., Heinrich, S.U., Graf, R., Nilsson, I., von Heijne, G., Brunner, J., and Rapoport, T.A. (1997). Molecular mechanism of membrane protein integration into the endoplasmic reticulum. Cell 89, 523–533. Nadal, M.S., Ozaita, A., Amarillo, Y., de Miera, E.V., Ma, Y., Mo, W., Goldberg, E.M., Misumi, Y., Ikehara, Y., Neubert, T.A., and Rudy, B. (2003). The CD26-related dipeptidyl aminopeptidase-like protein DPPX is a critical component of neuronal A-type K⫹ channels. Neuron 37, 449–461. Nagaya, N., and Papazian, D.M. (1997). Potassium channel alpha and beta subunits assemble in the endoplasmic reticulum. J. Biol. Chem. 272, 3022–3027. Nakatogawa, H., and Ito, K. (2002). The ribosomal exit tunnel functions as a discriminating gate. Cell 108, 629–636. Netzer, W.J., and Hartl, F.U. (1998). Recombination of protein domains facilitated by co-translational folding in eukaryotes. Nature 388, 343–349. Nicola, A.V., Chen, W., and Helenius, A. (1999). Co-translational folding of an alphavirus capsid protein in the cytosol of living cells. Nat. Cell Biol. 1, 341–345. Nilsson, T., Jackson, M., and Peterson, P.A. (1989). Short cytoplasmic sequences serve as retention signals for transmembrane proteins in the endoplasmic reticulum. Cell 58, 707–718. Nissen, P., Hansen, J., Ban, N., Moore, P.B., and Steitz, T.A. (2000). The structural basis of ribosome activity in peptide bond synthesis. Science 289, 920–930. O’Kelly, I., Butler, M.H., Zilberberg, N., and Goldstein, S.A. (2002). Forward transport. 14-3-3 binding overcomes retention in endoplasmic reticulum by dibasic signals. Cell 111, 577–588. Ottschytsch, N., Raes, A., Van Hoorick, D., and Snyders, D.J. (2002). Obligatory heterotetramerization of three previously uncharacterized Kv channel alpha-subunits identified in the human genome. Proc. Natl. Acad. Sci. USA 99, 7986–7991. Panyi, G., and Deutsch, C. (1996). Assembly and suppression of endogenous Kv1.3 channels in human T cells. J. Gen. Physiol. 107, 409–420. Patel, A.J., Lazdunski, M., and Honore, E. (1997). Kv2.1/Kv9.3, a

novel ATP-dependent delayed-rectifier K⫹ channel in oxygen-sensitive pulmonary artery myocytes. EMBO J. 16, 6615–6625. Paulussen, A., Raes, A., Matthijs, G., Snyders, D.J., Cohen, N., and Aerssens, J. (2002). A novel mutation (T65P) in the PAS domain of the human potassium channel HERG results in the long QT syndrome by trafficking deficiency. J. Biol. Chem. 277, 48610–48616. Pessia, M., Tucker, S.J., Lee, K., Bond, C.T., and Adelman, J.P. (1996). Subunit positional effects revealed by novel heteromeric inwardly rectifying K⫹ channels. EMBO J. 15, 2980–2987. Peterson, B.Z., DeMaria, C.D., Adelman, J.P., and Yue, D.T. (1999). Calmodulin is the Ca2⫹ sensor for Ca2⫹-dependent inactivation of L-type calcium channels. Neuron 22, 549–558. Potter, M.D., and Nicchitta, C.V. (2000). Regulation of ribosome detachment from the mammalian endoplasmic reticulum membrane. J. Biol. Chem. 275, 33828–33835. Rajamani, S., Anderson, C.L., Anson, B.D., and January, C.T. (2002). Pharmacological rescue of human K(⫹) channel long-QT2 mutations: human ether-a-go-go-related gene rescue without block. Circulation 105, 2830–2835. Salinas, M., Duprat, F., Heurteaux, C., Hugnot, J.P., and Lazdunski, M. (1997). New modulatory alpha subunits for mammalian Shab K⫹ channels. J. Biol. Chem. 272, 24371–24379. Sarma, J.D., Wang, F., and Koval, M. (2002). Targeted gap junction protein constructs reveal connexin-specific differences in oligomerization. J. Biol. Chem. 277, 20911–20918. Sato, Y., Sakaguchi, M., Goshima, S., Nakamura, T., and Uozumi, N. (2002). Integration of Shaker-type K⫹ channel, KAT1, into the endoplasmic reticulum membrane: synergistic insertion of voltagesensing segments, S3-S4, and independent insertion of pore-forming segments, S5-P-S6. Proc. Natl. Acad. Sci. USA 99, 60–65. Schulteis, C.T., Nagaya, N., and Papazian, D.M. (1998). Subunit folding and assembly steps are interspersed during Shaker potassium channel biogenesis. J. Biol. Chem. 273, 26210–26217. Schutze, M.P., Peterson, P.A., and Jackson, M.R. (1994). An N-terminal double-arginine motif maintains type II membrane proteins in the endoplasmic reticulum. EMBO J. 13, 1696–1705. Schwappach, B., Zerangue, N., Jan, Y.N., and Jan, L.Y. (2000). Molecular basis for K(ATP) assembly: transmembrane interactions mediate association of a K⫹ channel with an ABC transporter. Neuron 26, 155–167. Scott, D.B., Blanpied, T.A., Swanson, G.T., Zhang, C., and Ehlers, M.D. (2001). An NMDA receptor ER retention signal regulated by phosphorylation and alternative splicing. J. Neurosci. 21, 3063– 3072. Sharma, N., Crane, A., Clement, J.P., Gonzalez, G., Babenko, A.P., Bryan, J., and Aguilar-Bryan, L. (1999). The C terminus of SUR1 is required for trafficking of KATP channels. J. Biol. Chem. 274, 20628– 20632. Shen, N.V., Chen, X., Boyer, M.M., and Pfaffinger, P. (1993). Deletion analysis of K⫹ channel assembly. Neuron 11, 67–76. Shi, G., Nakahira, K., Hammond, S., Rhodes, K.J., Schechter, L.E., and Trimmer, J.S. (1996). Beta subunits promote K⫹ channel surface expression through effects early in biosynthesis. Neuron 16, 843–852. Skach, W.R. (2000). Defects in processing and trafficking of the cystic fibrosis transmembrane conductance regulator. Kidney Int. 57, 825–831. Standley, S., Roche, K.W., McCallum, J., Sans, N., and Wenthold, R.J. (2000). PDZ domain suppression of an ER retention signal in NMDA receptor NR1 splice variants. Neuron 28, 887–898. Stockklausner, C., Ludwig, J., Ruppersberg, J.P., and Klocker, N. (2001). A sequence motif responsible for ER export and surface expression of Kir2.0 inward rectifier K(⫹) channels. FEBS Lett. 493, 129–133. Teasdale, R.D., and Jackson, M.R. (1996). Signal-mediated sorting of membrane proteins between the endoplasmic reticulum and the golgi apparatus. Annu. Rev. Cell Dev. Biol. 12, 27–54. Tenson, T., and Ehrenberg, M. (2002). Regulatory nascent peptides in the ribosomal tunnel. Cell 108, 591–594.

Neuron 276

Thrift, R.N., Andrews, D.W., Walter, P., and Johnson, A.E. (1991). A nascent membrane protein is located adjacent to ER membrane proteins throughout its integration and translation. J. Cell Biol. 112, 809–821. Tiffany, A.M., Manganas, L.N., Kim, E., Hsueh, Y.P., Sheng, M., and Trimmer, J.S. (2000). PSD-95 and SAP97 exhibit distinct mechanisms for regulating K(⫹) channel surface expression and clustering. J. Cell Biol. 148, 147–158. Trudeau, M.C., and Zagotta, W.N. (2002). An intersubunit interaction regulates trafficking of rod cyclic nucleotide-gated channels and is disrupted in an inherited form of blindness. Neuron 34, 197–207. Tu, L., and Deutsch, C. (1999). Evidence for dimerization of dimers in K⫹ channel assembly. Biophys. J. 76, 2004–2017. Tu, L., Wang, J., Helm, A., Skach, W.R., and Deutsch, C. (2000). Transmembrane biogenesis of Kv1.3. Biochemistry 39, 824–836. Tucker, S.J., and Ashcroft, F.M. (1999). Mapping of the physical interaction between the intracellular domains of an inwardly rectifying potassium channel, Kir6.2. J. Biol. Chem. 274, 33393–33397. Tucker, S.J., Bond, C.T., Herson, P., Pessia, M., and Adelman, J.P. (1996). Inhibitory interactions between two inward rectifier K⫹ channel subunits mediated by the transmembrane domains. J. Biol. Chem. 271, 5866–5870. Wang, J.M., Zhang, L., Yao, Y., Viroonchatapan, N., Rothe, E., and Wang, Z.Z. (2002a). A transmembrane motif governs the surface trafficking of nicotinic acetylcholine receptors. Nat. Neurosci. 5, 963–970. Wang, X., Bogdanov, M., and Dowhan, W. (2002b). Topology of polytopic membrane protein subdomains is dictated by membrane phospholipid composition. EMBO J. 21, 5673–5681. Weitz, D., Ficek, N., Kremmer, E., Bauer, P.J., and Kaupp, U.B. (2002). Subunit stoichiometry of the CNG channel of rod photoreceptors. Neuron 36, 881–889. Wen, H., and Levitan, I.B. (2002). Calmodulin is an auxiliary subunit of KCNQ2/3 potassium channels. J. Neurosci. 22, 7991–8001. Xia, X.M., Fakler, B., Rivard, A., Wayman, G., Johnson-Pais, T., Keen, J.E., Ishii, T., Hirschberg, B., Bond, C.T., Lutsenko, S., et al. (1998). Mechanism of calcium gating in small-conductance calcium-activated potassium channels. Nature 395, 503–507. Xia, H., Hornby, Z.D., and Malenka, R.C. (2001). An ER retention signal explains differences in surface expression of NMDA and AMPA receptor subunits. Neuropharmacology 41, 714–723. Xu, J., Yu, W., Jan, Y.-N., Jan, L.-Y., and Li, M. (1995). Assembly of voltage-gated potassium channels. Conserved hydrophilic motifs determine subfamily-specific interactions between the alpha-subunits. J. Biol. Chem. 270, 24761–24768. Xu, J., Yu, W., Wright, J.M., Raab, R.W., and Li, M. (1998). Distinct functional stoichiometry of potassium channel beta subunits. Proc. Natl. Acad. Sci. USA 95, 1846–1851. Zarei, M.M., Zhu, N., Alioua, A., Eghbali, M., Stefani, E., and Stefani, L. (2001). A novel Maxi-K splice variant exhibits domainant-negative properties. Biophys. J. 80, 222a. Zerangue, N., Schwappach, B., Jan, Y.N., and Jan, L.Y. (1999). A new ER trafficking signal regulates the subunit stoichiometry of plasma membrane K(ATP) channels. Neuron 22, 537–548. Zerangue, N., Jan, Y.N., and Jan, L.Y. (2000). An artificial tetramerization domain restores efficient assembly of functional Shaker channels lacking T1. Proc. Natl. Acad. Sci. USA 97, 3591–3595. Zerangue, N., Malan, M.J., Fried, S.R., Dazin, P.F., Jan, Y.N., Jan, L.Y., and Schwappach, B. (2001). Analysis of endoplasmic reticulum trafficking signals by combinatorial screening in mammalian cells. Proc. Natl. Acad. Sci. USA 98, 2431–2436. Zheng, J., Trudeau, M.C., and Zagotta, W.N. (2002). Rod cyclic nucleotide-gated channels have a stoichiometry of three CNGA1 subunits and one CNGB1 subunit. Neuron 36, 891–896. Zhong, H., Molday, L.L., Molday, R.S., and Yau, K.W. (2002). The heteromeric cyclic nucleotide-gated channel adopts a 3A:1B stoichiometry. Nature 420, 193–198. Zhong, H., Lai, J., and Yau, K.-W. (2003). Selective heteromeric

assembly of cyclic nucleotide-gated channels. Proc. Natl. Acad. Sci. USA 100, 5509–5513. Zhou, Z., Gong, Q., Epstein, M.L., and January, C.T. (1998). HERG channel dysfunction in human long QT syndrome. Intracellular transport and functional defects. J. Biol. Chem. 273, 21061–21066. Zhou, Z., Gong, Q., and January, C.T. (1999). Correction of defective protein trafficking of a mutant HERG potassium channel in human long QT syndrome. Pharmacological and temperature effects. J. Biol. Chem. 274, 31123–31126. Zhu, X.R., Netzer, R., Bohlke, K., Liu, Q., and Pongs, O. (1999). Structural and functional characterization of Kv6.2, a new gammasubunit of voltage-gated potassium channel. Receptors Channels 6, 337–350.